(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Report of the President's Biomedical Research Panel : Supp 2"

Report of the 

President's 
Biomedical Research 

Panel 



SUPPLEMENT 2 

Impact of Federal Health-Related 

Research Expenditures Upon 

Institutions of Higher Education 



Uferaty 

Slstteal Institutes of IfeaUi 

Steyfcad 28914 



|& 4tf7/6-7. « 



•*'<&.. fe ^ 



y^y ' / 



Report of the 

President's 
Biomedical Research 

Panel 



SUPPLEMENT 2 

Impact of Federal Health-Related 

Research Expenditures Upon 

Institutions of Higher Education 

Studies by 

American Council on Education 

Association of American Medical Colleges 

The Rand Corporation 

(Contract Number NOl-PP-5-2159) 



April 30, 1976 



U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE 
U5. Public Health Service 
DHEW Publication No. (OS) 76-507 



7? 



$S3 
Sup. z 



For sale by the Superintendent of Documents, U.S. Government Printing Office 

Washington. D.C. 20402 - Price $6.65 

Stock No. 040-000-00352-5 






Preface 



The President's Biomedical Research Panel was established on January 29, 1975, and charged, 
under Public Law 93-352, to review and assess the conduct, support, policies, and management of 
biomedical and behavioral research as conducted and supported through programs of the National 
Institutes of Health (NIH) and the Alcohol, Drug Abuse, and Mental Health Administration 
(ADAMHA). The legislation directs the Panel to submit a report of its findings by April 30, 1976, 
to the President and to the Congress. 

Over the period of fifteen months, the seven members of the Panel conducted an extensive 
study that involved assessments of the state of the science, the impact of federally funded re- 
search on institutions of higher education, the organization and management of the NIH and the 
ADAMHA, the dissemination and application of research findings, and the development of policy 
for federal support of biomedical and behavioral research. 

The main volume of this report (DHEW Publication No. (OS)76-500) contains the issues iden- 
tified by the Panel in its study and the Panel's subsequent recommendations. The annex included 
in this volume details the methods of study used by the Panel and provides listings of witnesses 
who appeared before the Panel in hearings, professional and volunteer organizations contacted 
for submission of views, and resource materials available to the Panel in the course of its delibera- 
tions. Four appendices and four supplementary volumes accompany the main report. 

Appendix A, "The Place of Biomedical Science in Medicine and the State of the Science" 
(DHEW Publication No. (OS)76-501), contains the reports of the Overview and Interdisciplinary 
Clusters. In an unprecedented study, the Panel brought together preeminent biomedical and 
behavioral scientists and asked them, on an interdisciplinary basis, to assess the state of the 
science, to identify areas of promise and the resources required to achieve important new knowl- 
edge, and to relate research to the health of the nation's people. The results of this effort are 
contained within the reports of the eleven Interdisciplinary Clusters. The report of the Overview 
Cluster focuses on the place of biomedical science in medicine. 

Appendix B, "Approaches to Policy Development for Biomedical Research: Strategy for 
Budgeting and Movement from Invention to Clinical Application" (DHEW Publication No. 
(OS)76-502), presents the results of three studies performed under contract to the Panel: 
(1) "Lags Between Initial Discovery and Clinical Application to Cardiovascular Pulmonary Medi- 
cine and Surgery" (Julius H. Comroe, Jr.), (2) "Analysis of Selected Biomedical Research 
Programs" (BatteOe-Columbus Laboratories), and (3) "Policy Analysis for Federal Biomedical 
Research" (The Rand Corporation). 



in 



Appendix C, "Impact of Federal Health-Related Research Expenditures Upon Institutions of 
Higher Education" (DHEW Publication No. (OS)76-503), presents a summary of findings and 
conclusions of studies on the influence of federal funding of biomedical and behavioral research 
upon the financial status and educational functions of research universities and academic medical 
centers. The studies were conducted, under contract, by the American Council on Education, the 
Association of American Medical Colleges, and The Rand Corporation. 

Appendix D, "Selected Staff Papers" (DHEW Publication No. (OS) 76-5 04), offers eight papers 
prepared by individual members of the Panel staff on a variety of topics pertinent to the NIH 
and the ADAMHA. 

All five of the above mentioned volumes are available from the Government Printing Office, 
as indicated. 

In addition, four volumes of supplementary resource documents are also available from the 
Government Printing Office. 

Supplement 1. "Analysis of Selected Biomedical Research Programs: Case Histories" (DHEW 
Publication No. (OS)76-506). Background information for the study by 

Battelle-Columbus Laboratories. 

Supplement 2. "Impact of Federal Health-Related Research Expenditures Upon Institutions 
of Higher Education" (DHEW Publication No. (OS)76-507). Findings, con- 
clusions, and background information of studies by the American Council on 
Education, the Association of American Medical Colleges, and The Rand 
Corporation. 

Supplement 3. "Written Statements Supplementing Verbal Testimonies of Witnesses" (DHEW 
Publication No. (OS)76-508). 

Supplement 4. "Statements of Professional, Scientific, and Voluntary Health Organizations" 
(DHEW Publication No. (OS)76-509). 

Verbatim transcripts for all meetings of the Panel, except the March 1975 meeting, are avail- 
able from the U.S. Department of Commerce, National Technical Information Service, 
Springfield, Virginia 22161. 



IV 



CONTENTS 



Part I. A Study of Financial and Educational Trends in Universities 
in Relation to Federal Funding of Health-Related Research 

American Council on Education 



Part II. Effects of Federal Funds Upon Selected Health-Related 
Disciplines 

The Rand Corporation 



Part III. Trends and Dimensions of Biomedical and Behavioral Research 
Funding in Academic Medical Centers 

Association of American Medical Colleges 



Part IV. The Effect of Federal Biomedical Research Programs on 
Academic Medical Centers 

The Rand Corporation 



Part V. A Report on the Indirect Costs of Academic Research 

American Council on Education 
Association of American Medical Colleges 
The Rand Corporation 



Part VI. Indirect Costs in Universities 

American Council on Education 



PART I 



A Study of Financial and Educational Trends in Universities 
in Relation to Federal Funding of Health-Related Research 

1964-1974 



Lyle H. Lanier and Ivars Zageris 



American Council on Education 



TABLE OF CONTENTS 

Page 

-L • J. IN J. r\LJlJU ^lX*^/lN[?POOO«»00»OOOeOOO»»000000 000000 00000 30QOOOOO _L 

II . PROCEDURE 4 

III. TRENDS IN TOTAL EDUCATIONAL -AND-GENERAL REVENUES 13 

IV. TRENDS IN REVENUES FOR SPONSORED RESEARCH 21 

V. TRENDS IN EXPENDITURES FOR BIOMEDICAL-BEHAVIORAL 

RESEARCH 33 

VI . TRENDS IN UNIVERSITY ENROLLMENTS . 49 

VII . TRENDS IN EARNED DOCTORAL DEGREES 64 

VIII. TRENDS IN THE FUNDING PATTERNS OF NIH AND ADAMHA 75 

IX. SUMMARY AND CONCLUSIONS 86 

REFERENCES 96 

APPENDIX A - Classification of the 148 Institutions in 
the ACE-RAND Sample by the Categories of the 
Carnegie Commission on Higher Education 97 

APPENDIX B - Tables Showing Expenditures for Biomedical- 
Behavioral Research for Institutions Classified by 
Categories of the Carnegie Commission on Higher 
Education 100 

APPENDIX C - Enrollment Tables for Institutions Classi- 
fied by Categories of the Carnegie Commission on 
Higher Education 107 

APPENDIX D - Tables Showing Earned Doctoral Degrees for 
Institutions Classified by Categories of the Carne- 
gie Commission on Higher Education 114 



■x- 



LIST OF TABLES 



Text Tables 



Table 1. Questions Raised by the President's Biomedical Research 

Panel Regarding Dimensions and Trends in Research Funding 
in Universities 



Page 



Table 2. Distribution of the Numbers of Institutions in the ACE 
Sample by (a) Carnegie Commission Classification, (b) 
Presence or Absence of a Medical School, and (c) Type of 
Control 

Table 3. Classification of Types and Sources of Data Included in 

the ACE Data Base for Universities 

Table 4. Price Deflators for Four Types of Higher-Education Expen- 
ditures, together with the Consumer Price Index and the 
Wholesale Price Index 

Table 5. Trends in Mean Educational-and-General Revenues by Type 
of Institution — Constant Dollars in Thousands (Halstead 
HEPI as Deflator , FY 1964 = 100) 

Table 5A. Trends in Index Numbers for the Means of Educational-and- 
General Revenues Shown in Table 5 (Means for FY 1972 = 
100) 



12 



14 



14 



Table 6. Trends in Mean Educational-and-General Revenues by 

Carnegie Commission Categories of Institutions — Constant 
Dollars in Thousands (Halstead HEPI as Deflator, FY 1964 
= 100) 

Table 6A. Trends in Index Numbers for the Means of Educational-and- 
General Revenues Shown in Table 6 (Means for FY 1972 = 
100) 

Table 7. Median Educational-and-General Expenditures per Full- 
Time-Equivalent Student by Carnegie Commission Cate- 
gories (Constant Dollars) 

Table 7A. Median Percentage Change in Educational-and-General 
Expenditures per FTE Student by Carnegie Commission 
Categories (in Current and in Constant Dollars) 

Table 8. Trends in Total and in Federally Sponsored R&D Revenues 

as Percentages of Educational-and-General Revenues (Based 
on Constant Dollars) 



17 



17 



19 



19 



23 



-ii- 



Text Tables (cont.) Page 

Table 9. Trends in Mean Sponsored R&D Revenues by Type of Insti- 
tution — Constant Dollars in Thousands (NIH R&D Deflator, 
FY 1964 = 100) 26 

Table 9A. Trends in Index Numbers for the Means of Sponsored R&D 

Revenues Shown in Table 9 (Means for FY 1972 = 100) 26 

Table 10. Trends in Mean Sponsored R&D Revenues by Carnegie Comm- 
ission Categories of Institutions — Constant Dollars in 
Thousands (NIH R&D Deflator, FY 1964 = 100) 28 

Table 10A. Trends in Index Numbers for the Means of Sponsored R&D 

Revenues Shown in Table 10 (Means for FY 1972 = 100) 28 

Table 11. Trends in Mean Federally Sponsored R&D Revenues by Type 
of Institution — Constnat Dollars in Thousands (NIH R&D 
Deflator, FY 1964 = 100) 30 

Table 11A. Trends in Index Numbers for the Means of Federally Spon- 
sored R&D Revenues Shown in Table 11 (Means for FY 1972 
= 100) 30 

Table 12. Trends in Mean Federally Sponsored R&D Revenues by 
Carnegie Commission Categories — Constant Dollars in 
Thousands (NIH R&D Deflator, FY 1964 = 100) 32 

Table 12A. Trends in Index Numbers for the Means of Federally Spon- 
sored R&D Revenues Shown in Table 12 (Means for FY 1972 
= 100) 32 

Table 13. Trends in Mean Expenditures for All Biomedical-Behav- 

ioral Research by Type of Institution — Constant Dollars 

in Thousands (NIH R&D Deflator, FY 1964 = 100) 36 

Table 13A. Trends in Index Numbers for the Means of All Expenditures 
for Biomedical-Behavioral Research Shown in Table 13 
(Means for FY 1972 = 100) 36 

Table 14. Trends in Mean Federally Funded Expenditures for Biomedi- 
cal Behavioral Research by Type of Institution — Constant 
Dollars in Thousands (NIH R&D Deflator, FY 1964 = 100)..... 38 

Table 14A. Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Biomedical-Behavioral Research Shown in 
Table 14 (Means for FY 1972 = 100) 38 



Table 15. Trends in Mean Federally Funded Expenditures for Bio- 
logical (Agriculture Included) Research by Type of Insti- 
tution — Constant Dollars in Thousands (NIH R&D Deflator, 
FY 1964 = 100) 41 

Table 15A. Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Biological Research Shown in Table 15 
(Means for FY 1972 = 100) 41 



-in- 



Text Tables (cont.) 

Table 16. Trends in Mean Federally Funded Expenditures for Medical 
Research by Type of Institution — Constant Dollars in 
Thousands (NIH R&D Deflator, FY 1964 = 100) 

Table 16A. Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Medical Research Shown in Table 16 
(Means for FY 1972 = 100) 

Table 17. Trends in Mean Federally Funded Expenditures for Research 
in Life Sciences Not Elsewhere Classified, by Type of 
Institution — Constant Dollars in Thousands (NIH R&D 
Deflator, FY 1964 = 100) 

Table 17A. Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Research in Life Sciences Not Elsewhere 
Classified Shown in Table 17 (Means for FY 1972 = 100)... 

Table 18. Trends in Mean Federally Funded Expenditures for Psycho- 
logy by Type of Institution — Constant Dollars in Thou- 
sands (NIH R&D Deflator , FY 1964 = 100) 

Table 18A. Trends in Index Numbers for the Means of Federally Funded 
Research in Psychology Shown in Table 18 (Means for FY 
1972 = 100) 

Table 19. Trends in Mean Degree-Credit Enrollment by Type of 

Institution 

Table 19A. Trends in Index Numbers for the Means of Degree-Credit 

Enrollment Shown in Table 19 (Means for FY 1972 = 100) . . . 

Table 20. Trends in Mean Enrollment for Advanced Degrees in All 

Fields by Type of Institution. 

Table 20A. Trends in Index Numbers for the Means of Enrollment for 
Advanced Degrees in All Fields Shown in Table 20 (Means 
for FY 1972 = 100) 

Table 21. Trends in Mean Enrollment for Advanced Degrees in Bio- 

medical-Behavioral Sciences by Type of Institution 

Table 21A. Trends in Index Numbers for Means of Enrollment for 

Advanced Degrees in Biomedical-Behavioral Sciences Shown 
in Table 21 (Means for FY 1972 = 100) 

Table 22. Trends in Mean Enrollment for Advanced Degrees in Bio- 
logical Sciences by Type of Institution 

Table 22A. Trends in Index Numbers for the Means of Enrollment for 
Advanced Degrees in Biological Sciences Shown in Table 
22 (Means for FY 1972 = 100) 



Page 

43 
43 

45 
45 

47 

47 

51 
51 

54 

54 
57 

57 
59 

59 



-iv- 



Text Tables (cont.) Page 

Table 23. Trends in Mean Enrollment for Advanced Degrees in Health 

Professions by Type of Institution 61 

Table 23A. Trends in Index Numbers for the Means of Enrollment for 
Advanced Degrees in Health Professions Shown in Table 23 
(Means for FY 1972 = 100) 61 

Table 24. Trends in Mean Enrollment for Advanced Degrees in Psycho- 
logy by Type of Institution 63 

Table 24A. Trends in Index Numbers for the Means of Enrollment for 
Advanced Degrees in Psychology Shown in Table 24 (Means 
for FY 1972 = 100) 63 

Table 25. Trends in Mean Number of Earned Doctoral Degrees in All 

Fields by Type of Institution 65 

Table 25A. Trends in Index Numbers for the Means of Earned Doctoral 
Degrees in All Fields Shown in Table 25 (Means for FY 
1972 = 100) 65 

Table 26. Trends in Mean Number of Earned Doctoral Degrees in Bio- 

medical-Behavioral Sciences by Type of Institution........ 68 

Table 26A. Trends in Index Numbers for the Means of Earned Doctoral 
Degrees in oiomedical-Behavioral Sciences Shown in Table 
26 (Means for FY 1972 = 100) 68 

Table 27. Trends in Mean Number of Earned Doctoral Degrees in Bio- 
logical Sciences by Type of Institution. 71 

Table 27A. Trends in Index Numbers for the Means of Earned Doctoral 
Degrees in Biological Sciences Shown in Table 27 (Means 
for FY 1972 = 100) 71 

Table 28. Trends in Mean Number of Earned Doctoral Degrees in Health 

Professions by Type of Institution . 72 

Table 28A. Trends in Index Numbers for the Means of Earned Doctoral 
Degrees in Health Profession Shown in Table 28 (Means 
for FY 1972 = 100) 72 

Table 29. Trends in Mean Number of Earned Doctoral Degrees in Psy- 
chology by Type of Institution 73 

Table 29A. Trends in Index Numbers for the Means of Earned Doctoral 
Degrees in Psychology Shown in Table 29 (Means for FY 
1972 = 100) 73 



-v- 



Text Tables (cortt.) Page 

Table 30. Trends in Percentage Distributions for NIH Awards by 

Funding Mechanism to 145 Universities 77 

Table 31. Trends in Percentage Distributions for ADAMHA Awards by 

Funding Mechanism to 145 Universities 77 

Table 32. Trends in Mean NIH Awards to 145 Universities by Funding 
Mechanism — Constant Dollars in Thousands (NIH R&D Defla- 
tor, FY 1964 = 100) 79 

Table 32A. Trends in Index Numbers for the Means of NIH Awards Shown 

in Table 32 (Means for FY 1972 = 100) 79 

Table 33. Trends in Mean ADAMHA Awards to 145 Universities by 

Funding Mechanism — Constant Dollars in Thousands (NIH 

R&D Deflator , FY 1964 = 100) 81 

Table 33A. Trends in Index Numbers for the Means of ADAMHA Awards 

Shown in Table 33 (Means for FY 1972 = 100) 81 

Table 34. Trends in Direct and Indirect Costs as Percentages of 

Total NIH Awards to 145 Universities by Funding Mechanism 82 

Table 35. Trends in Direct and Indirect Costs as Percentages of 
Total ADAMHA Awards to 145 Universities by Funding 
Mechanism 83 



-vi- 



I. INTRODUCTION 

The general purpose of this study has been to investigate dimensions and 
trends in federal funding of biomedical-behavioral research in universities 
over the period from 1963-64 through 1973-74 in relation to concurrent trends 
in the overall financial and educational operations of these institutions. Spon- 
sored by the President's Biomedical Research Panel, it is part of a more compre- 
hensive project on the impacts of health-related research expenditures upon the 
financial status, instructional programs, faculties, and educational outputs of 
universities and academic medical centers. 

The Panel defined the specific nature and scope of its interest in the 
present study in the form of a set of six questions, which are reproduced in 
Table 1 on the following page. They fall generally into two groups: (a) ques- 
tions pertaining to financial trends and their interrelationships in research 
universities, with special reference to the funding of health-related research; 
(b) questions concerned with concurrent trends in the educational resources and 
activities of these institutions. 

The American Council on Education has undertaken a comprehensive program 
of statistical investigation in an effort to answer these specific questions 
and, insofar as available data would permit, to address the more general issues 
raised in the Panel's background statement of its premises and purposes. A 
sample of institutions was first established which included all U.S. universi- 
ties with medical schools and a matching group without medical schools. Together 
they comprise all institutions classified by the Carnegie Commission on Higher 
Education as "research universities" and most of the remainder that award doc- 
toral degrees. Data were then sought for the institutional sample that would 



-2- 



Table 1 . Questions Raised by the President's Biomedical Research Panel Regarding 
Dimensions and Trends in Research Funding in Universities 



1. What is the total educational and general revenue of universities? 

2. What part of the above is provided for: 

a. sponsored research (all funding sources)? 

b. federal R&D projects (all agencies)? 

c. federal biomedical and behavioral research projects? 

d. non-federal R&D projects? 

3. What part of funding of new construction and renovation has been provided 
by federal sources? Non-federal sources? What part of each of the pre- 
ceding has been for biomedical and behavioral research? 

4. What trends in funding have occurred with the following instruments for 
the transfer of NIH and ADAMHA funds to universities through: 

a. regular research grants? 

b. program project grants? 

c. center grants? 

d. contracts? 

e. training grants? 

f. faculty awards? 

g. general research grants? 
h. clinical research centers? 

i. construction and renovation grants and loans? 
j . National Library of Medicine awards? 
k. other? 

5. To the extent that data are available from other federal agencies, similar 
breakouts as they apply to biomedical and behavioral research shall also 
be studied. 

6. Between the years 1964 through 1974, what changes have there been in 
faculties and students in: 

a. total student enrollment? 

b. graduate student enrollment? 

c. biomedical and behavioral student enrollment? 

d. number of biomedical and behavioral postdoctoral students? 

e. number of graduate assistants? 

f. graduate degrees granted, by fields? 



-3- 



permit the kinds of analyses required to provide answers to the Panel's ques- 
tions. Because of the project's original duration (eight months), however, it 
was necessary to limit the data base to statistical information already avail- 
able in the files of federal and private agencies. 

It turned out that these sources either had no data or entirely inadequate 
data relative to certain of the questions. Furthermore, even when the types 
of data appeared to be satisfactory, most of the files were deficient in one 
or more of the following respects: (a) failure to cover the full period under 
study (FY 1964 through FY 1974) ; (b) incompatibility between earlier and later 
records in a series due to questionnaire changes (e.g., varying definitions of 
an item); (c) changes in the institutional composition of the reporting units 
(mainly in the case of multicampus universities) ; (d) institutional data missing 
for one or more of the years covered by a given data series; (e) inconsisten- 
cies among sources in reporting data for ostensibly the same variable. 

Despite these limitations, it was possible to discover significant trends 
and interrelations among the data that were available and usable. Such findings 
include differences in several dimensions between various subdivisions of the 
total institutional sample: private vs. public control; presence vs. absence 
of a medical school; the hierarchical categories developed by the Carnegie 
Commission on Higher Education. 

The questions in Table 1 for which no relevant trend data could be found 
were the following: (a) item 3 on the sources and purposes of funding for new 
construction and renovation; (b) item 5 on trends in federal support by funding 
mechanisms for agencies other than NIH and ADAMHA; (c) numbers of faculty mem- 
bers, graduate assistants, and postdoctoral students by discipline. 



-4- 



II. PROCEDURE 

Three general types of procedures will be described: (a) selection of 
the sample of universities; (b) construction of the data base; (c) analytical 
techniques and modes of presenting results. 

Sample of Institutions 

As a point of departure for selecting the ACE sample, it was decided to 
include all university campuses with medical schools within their organizational 
structure and jurisdiction (whether or not physically located on the campus) . 

In selecting a parallel group of universities without medical schools, the 
initial intention was to limit them to the top three categories of institutions 
in the classification hierarchy devised by the Carnegie Commission on Higher 
Education: Research Universities I, Research Universities II, and Doctoral 
Universities I. Some 87 per cent of all universities with medical schools on 
campus (as defined above) fell within these three groups, and the few institu- 
tions with medical schools not classified among the top three categories appeared 
to be difficult to "match" precisely with "non-medical-school" counterparts. 
But it was finally decided to include four public campuses without medical 
schools from other categories — mainly because they were members of multi-campus 
institutions which earlier had reported only aggregated data for all its cam- 
puses, and it seemed desirable to maintain the continuity of such aggregates 
for possible use in certain of the analyses. 



-5- 



The top five categories of the classification developed by the Carnegie 

Commission, from which the sample for Task 1 was drawn, are defined in terms 

of the following criteria: 

Research Universities I . The 50 leading universities in terms of federal 
financial support of academic science in at least two of the three acade- 
mic years, 1968-69, 1969-70, and 1970-71, provided they awarded at least 
50 Ph.D.'s (plus M.D.'s if a medical school was on the same campus) in 
1969-70. Rockefeller University was included because of the high quality 
of its research and doctoral training, although it did not meet these 
criteria. 

Research Universities II . These universities were on the list of the 100 
leading institutions in terms of federal financial support in at least 
two out of the above three years and awarded at least 50 Ph.D.'s (plus 
M.D.'s if a medical school was on the same campus) in 1969-70, or they 
were among the leading 50 institutions in terms of the total number of 
Ph.D.'s (plus M.D.'s if on the same campus) awarded during the years from 
1960-61 to 1969-70. In addition, a few institutions that did not quite 
meet these criteria, but that have graduate programs of high quality and 
with impressive promise for future development, have been included in 
Research Universities II. 

Doctoral-Granting Universities I . These institutions awarded 40 or more 
Ph.D.'s in 1969-70 (plus M.D.'s if on the same campus) or received at 
least $3 million in total federal financial support in either 1969-70 or 
1970-71. No institution is included that granted fewer than 20 Ph.D.'s 
(plus M.D.'s if on the same campus), regardless of the amount of federal 
financial support it received. 

Doctoral-Granting Universities II . These institutions awarded at least 
10 Ph.D.'s in 1969-70, with the exception of a few new doctoral-granting 
institutions that may be expected to increase the number of Ph.D.'s 
awarded within a few years. 



Carnegie Commission on Higher Education. A Classification of Institutions of 
Higher Education . A Technical Report. Berkeley, Calif.: Carnegie Commission 
on Higher Education, 1973. 

2 
The term "Ph.D." in these definitions includes the Ed.D. and other doctoral 

degrees (but not M.D.'s, D.D.S.'s etc.). 



-6- 



Comprehensive Universities and Colleges I . This group includes institu- 
tions that offered a liberal arts program as well as several other programs, 
such as engineering and business adminstration. Many of them offered 
master's degrees, but all lacked a doctoral program or had an extremely 
limited doctoral program. All institutions in this group had at least 
two professional or occupational programs and enrolled at least 2,000 
students in 1970. If an institution's enrollment was smaller than this, 
it was not considered comprehensive. 

The "three-dimensional" distribution of the numbers of institutions in the 
sample is presented in Table 2 on the following page: (a) by Carnegie Commis- 
sion category; (b) by presence or absence of a medical school; and (c) by type 
of control (private, public). The last section of the table shows the corres- 
ponding numbers of universities in the national population of such institutions, 
with breakdowns by type of control. A comparison between the sample and the 
population totals shows that the original intention to include in the sample 
all institutions from the top three Carnegie Commission categories was not 
quite realized. Missing are three of the 52 Research Universities I and six 
of the 53 Doctoral Universities I. Of the missing nine institutions, two 
are private and seven are public. All omissions were due to excessive incom- 
pleteness or unusability of records. 

The sample of 148 universities received more than 80 per cent of all 
federal funds obligated for research and development according to the National 
Science Foundation's data for FY 1972. Specific percentages were: total R&D 
for all fields, 83%; total R&D for life sciences, 81%; total R&D for psychology, 
84%. 



-7- 








01 




01 


i-( 


•H 


O 


£ 


in 


£ 


4-1 





c 


u 


o 




o 


cu 




•H 


KH 


M 


O 


0) 




c 


111 


M 


o. 


co 


>^ 


U 


H 



(0 o 



>* XI 


XI 


c 




CO 


0) 




r-i 


•1 


P, 


rH 


s 


o 


co- 


o 


co 


J= 




o 


w 


CO 


cj 




< 


i-H 




CO 


cu 


o 


x: 


•H 


j-j 


XI 




11 


c 


s 


•H 






cd 


oi 




C 


•IH 





o 


-H 




4.) 


(U 


3 


cj 


u 


c 


•H 


m 


4J 


0) 


7] 


XI 


C 


< 


l-l 






1-1 


■4H 


o 









a' 


in 


o 


>-i 


c 


0) 


a 


XI 


w 


1 


CU 
m 


z 


Ph 


OJ 


/— V 


J= 


XI 


4-J 


^^ 


«H 


• 


o 


c 




o 


c 


•H 


o 


4J 


T-i 


CO 


4J 


CJ 


— j 


i-l 


.5 


•W 


iH 


•H 


in 


7] 


4-1 


CO 


71 


CO 



a o 



01 


c 


CC 


x: 


o 


4-1 


4-1 


•H 


O 




4-1 


H 


c 


CC 




-r4 


■H 


o 




3 


•H 


CO 


c^ 


rH 


c 





XI 


o 


Ph 


3 


•H 




Pm 


4-1 


r-H 




3 


CO 


OJ 


4J 


c 


4-J 


•H 


c 


CO 


4-1 


•H 


> 


oi 


4-1 


•H 


C 


CO 


t-l 


i-i 


f. 


P-. 



CM 
6 

cd 

C/l 

W 
U 

0) 

x: 



CO 




c 


0) 


o 


.-1 


H 


a. 


4-J 


E 


3 


CO 


4-1 


CO 


H 




1-1 


OJ 


71 


X 


C 


4-1 




a 


i— 1 
3 


■H 



J30 


4J 


0) 


•H 


H 


s 


O 




o 


CO 


x; 


c 


u 


o 


lo 


■H 




4-> 


rH 


3 


CO 


4-1 


u 


•H 


iH 


4-J 


-a 


01 


O) 


c 


S3 


1-1 





c 









i-W 


CO 


c 


71 


o 


•H 


•H 


E 


4-J 


3 


CO 


O 


CJ 


CJ 


•r-t 




•4H 


eu 


•H 


•H 


71 


00 


7) 


0) 


CO 


c 


rH 


m 


CJ 


CO 




CJ 





CM o PI 

m <r in 



O r~ ■* 
co cm CO 



cm ro en 

CM t-H t-H 



cr> o r-» 

•* <f •* 



r^ r^ O 
cm cm co 



cm co r-~ 
CN iH H 



co \0 r^ 

r-l CM CO 



o> o"\ in 

t-H CM 



-3- r~- cm 



O 
CO 



-* 


CT\ 


CO 


O 


CO 


m 



CO 



O 



vO -3- O CO 

CO rH rH 



oo oo in co 



co •o m o 



CO 
CO 



•J3 
•43 



CO O0 



in co 



co m 
uo 



CO o 

00 



O CO 
CM 



co 











CO 














01 








M 




M 


i-l 






M 


M 


M 


M 


4-1 
•H 






CO 


CO 


CO 


co 


CO 






cu 


01 


01 


01 


I-l 






1-1 


i-t 


•H 


•H 


a) 






4-J 


4-1 


4-1 


4-1 


> 


M 




•H 


•H 


i-l 


•H 


•H 






co 


CO 


co 


0) 


c 


CO 




)H 


U 


u 


M 


3 


01 




0) 


01 


01 


01 




00 




> 


> 


> 


> 


01 


01 




1-1 


■H 


•H 


•H 


> 


.-1 




c 


c 


C 


c 


1-1 


i-H 




3 


D 


D 


o 


co 

c 


o 


■-I 


x; 


x: 


^4. 


t-l 


01 




cd 


CJ 


o 


CO 


tfl 


X! 


T3 


4-J 


l-l 


M 


I-l 


u 


01 


c 


o 


cd 


td 


o 


o 


|4 


cs 


H 


at 


01 


4-J 


4-J 


a. 






CO 


CO 


O 


CJ 


i- 






01 


01 





o 


o 






Pi 


pi 


o 


a 


u 







01 

c 
o 



OJ 




•H 








CO 




00 




4-1 








3 




0> 




3 








P. 




T-) 




4-J 








6 




rH 




CO -H 








co 




o 




a) 4-J 








CJ 




CJ 




00 CO 

01 C 








cu 




•4-1 




r-l M 








XI 




o 




rH 

O T3 






. 


4-1 




c 




CJ 0) 






4-1 


IH 




O 




N 






X 


01 




•H 




IH -H 






01 


TD 




4-J 




Cd rH 






4-J 


C 




cd 




01 CO 








3 




CJ 




>H . H 






0) 






•H 




1 CJ 






rS 


01 




14-1 




O 01 






4_1 


71 




•H 




3 p- 








rH 




71 




H CO 






c 


OJ 




71 










•H 






cd 












IH 




i-H 




» 






T3 


o 




O 




• O 
CTi rH 






0> 
XI 


71 




CO 










•H 


3 




4-1 










Sh 


Cu 




•H 




M 






o 

71 


E 
cd 




C 




l-H l-H 






01 


CJ 




•H 




CO 71 






TD 


01 




71 




01 01 






0) 


£- 




0) 




00 00 






^4 


co 




•H 




01 01 






cd 


71 




1-4 




rH rH 












O 




rH rH 






01 


0) 




00 




O O 




1—1 


^ 


rC 




0> 




CJ C_> 


M 


rH 


Cu 


4-J 




4-J 










B 






CO 




•41 •£ 


01 


71 


fl 


a 




CJ 




« • 


0) 

oo 


0) 
043 


CO 


o 




00 




> > 


a) 


QJ 


w 


XI 




c 




•H iH 


rH 


t— 1 


CJ 


01 




•H 




c c 


rH 


rH 


< 


4-1 




3 




P D 


O 


O 




CO 




o 






CJ 


CJ 


cu 


CJ 




.-1 




CU 0) 






x; 


o 




rH 




> > 


CO 


71 


4J 


rH 




O 




•H iH 


4-1 


4J 








•4-1 




CO CO 


I-l 


I-l 


c 


CO 








c c 


< 


< 


■H 


rH 




01 




01 01 








O 




x; 




x: x; 


rH 


rH 


■d 


o 




4J 




01 01 


CO 


cfl 


0) 


X! 








M I-l 


I-l 


1-4 


CO 


CJ 




T3 




a a. 


01 


CD 


3 


CO 




0) 




e e 


Xl 


Xi 








■a 




o o 


■H 


■H 


co 


rH 




3 




o o 


rJ 


r-) 


cu 


CO 




r-l 










•rl 


CJ 




CJ 










IH 


■H 




C 




• • 


• 


• 


O 


•v 




•H 




m \o 


r-~ 


00 


00 
01 


0) 

E 




c 










4J 






o 










cd 


x: 




•H 




M 




rH 


CJ 


4J 




4-J 




M rH 


l-H 


M 




•H 




cd 










01 


3 




CJ 




co co 


CO 


71 


> 






3 




01 01 


01 


01 


•H 


CO 




TJ 




■H vH 


•H 


■H 


•4H 


C 




W 




4-1 J-i 


4-1 


4H 




o 








•H -rl 


•H 


•H 


01 


•r-l 




l-i 




CO CO 


01 


01 


XI 


4-1 




01 




rl I-l 


>-! 


IH 


4-J 


3 




x: 




0) 0' 


cu 


aj 




4-J 




00 




> > 


> 


> 


u 


•H 




•H 




iH iH 


•H 


•H 


o 


4-1 




3= 




C C 


c 


C 


<4H 


co 








3 D 


3 


e> 




C 




c 










CO 


•H 




o 




x: x- 


H 


rH 


iH 










u cj 


CO 


« 


IH 


>> 




c 




IH IH 


IH 


1-1 


0! 


t-H 




o 




CO CO 


o 


o 


4J 


C 




•H 




OJ 01 


JJ 


4J 


•H 


O 




CO 




CO CO 


o 


CJ 


IH 






CO 




01 OJ 


o 


o 


O 


CO 




•H 




OS Pi 


D 


p 




01 




§ 










C 


XI 




e 


•« 








O 


3 




o 


71 


• ■ 


• 


■ 


•H 


rH 




CJ 


CD 

•t-l 


rH CM 


CO 


<r 


4-1 
CO 


CJ 

c 




01 


4-1 








CJ 


-H 


• 


•H 


•t-l 








•H 




C 


oo 


(0 








•4H 


a 


o 


01 


>-l 








•H 


3 


•H 


c 


OJ 








CO 


o 


4H 


M 


> 








co 


IH 


CJ 


CO 


•H 








cd 


00 


•H 


O 


c 








rH 




X 




3 








CJ 


01 


01 


01 












•H 


•H 


J3 


X) 








0) 


X! 


IH 


H 


C 








X! 


H 


3 


cd 


CO 








H 


Xl 


•n 



-8- 



Composition of the Data Base 

As noted in the Introduction, the information used in this study was 
limited to data already existing in the files of various federal agencies and 
private organizations. Two general types of data were assembled: financial 
statistics; educational and personnel statistics. The number of data elements 
in both categories totaled more than 150, which were compiled from computer 
tapes and hard-copy records provided by the following sources: 

1. National Center for Education Statistics (NCES) 

2. National Science Foundation (NSF) 

3. National Institutes of Health (NIH) 

4. Alcohol, Drug Abuse, and Mental Health Administration (ADAMHA) 

5. National Academy of Sciences-National Research Council (NAS-NRC) 

6. American Association of University Professors (AAUP) 

The general categories of data supplied by these organizations are indica- 
ted in Table 3 on the following page, which is divided into two sections: (a) 
financial statistics, from three agencies; (b) non-financial statistics, from 
four agencies. 

With 148 institutions in the sample, if each one had supplied all of the 
desired data elements for the 11 years covered by the study the data base would 
have had a total of about a quarter of a million statistical items. Unhappily, 
the actual number fell far short of that theoretical figure for reasons already 
cited: (a) most of the surveys did not span the entire 11-year period, and 
some that did were conducted intermittantly; (b) in certain surveys, changes 
in the definition of the data elements occurred, including the level of aggrega- 
tion; (c) many institutions did not supply all of the data requested; (d) changes 
in the composition of the reporting unit in the case of several multicampus 
institutions (e.g., aggregate data for all campuses in a system were reported 
for certain years and for other years reports by individual campuses were sub- 
mitted) . 



-9- 

















CO 


CD 
















X) 


4-1 


SH 
















rH 


3 


o 






43 








*—S 


QJ 


CD 


CD 




CD 


CJ 








CD 


•H 


•H 


CD 




S-i 


cd 






CD 


X) 


PH CD 


& 


CD 




CD 


CD 









rH 


Xl 


•H 


m 




& 








•rl 


CD 


>^ rH 


CJ 


o 






3 






4-1 


•H 


43 0) 


CD 


S-l 




CO 


•H 






CD 


Ph 


•H 


Pd 


fc 




CD 








•H 




CO Ph 








rH 


X) 






4J 


>^ 


U 


rH 


^ 




43 


Q) 






cd 


42 


CD >. cd 


4-1 




cd 


42 






4J 




a) -a 


SH 44 


•H 




•H 


•H 






Wf 


1 CD 


3 


O S-l 


CD 




M 


S-l 








44 CD 


•H CO 


CD 4J O 


S-l 




cd 


CJ 






3 


Cd XI 


60 4J 


tj o a- 


CD 




> 


CD 






o 


3 rH 


3 3 


rH O D, 


> 


t^4*! 




0) 






•H 


T3 CD 3 


w cd 


ai q 3 


•H 


44 3 


CD 


XI 






4-1 


cd -H o 


4-1 


•H CO 


3 


rH Cd 


Xi 








cd 


M Pn -H 


X) CD 


rH 


Ph mh 


C3 


3 Pd 


44 


CD 




CD 


CJ 


O 4-1 


3 -H 


•H 


O ^ 




CJ 




43 




CJ 


3 


>% cd 


cd co 


CJ 


r^ cd 


4H 


cd >> 


MH 






■H 


XI 


* rO TJ 


CO 


3 


,n co Sh 


O 


Ph 43 


o 


^-i 




4J 


W 


rH 3 


CO <i 


3 


3 








rH 




CD 




cd cd 3 


4J 


O 


CO Cd 4-J 


3 


CD CO 


0) 


•H 




•H 


U 


44 CD O 


CD CD 


CJ 


CD rH O 


O 


e <D 


6 


& 


"3 


44 


O 


O CD Ph 


•H 4J 




CD PH O 


•H 


•H -H 


o 




CD 


ca 


MH 


H H 


4-1 cd 


43 


S-l X) 


4-1 


H H 


CD 


XI 


-d 


4J 




^-' 00 CD 


3 = 


o 


00 rH 4-1 


cd 


i cd 




CD 


3 


CO 


SH 


CD CJ 


CD 13 


Sh 


cd cd co 


•H 


r-\ r-1 


•s 


CO 


H 




CD 


4-1 Q 3 


•H cd 


cd 


O Sh O 


a 


r-i cd 


XI 


3 





H 


4-1 


3 CD 


CJ Sh 


CD 


O Ph 


o 


3 CO 


CD 




c 


cd 


a 


CD rH -H 


CO o 


CO 


rH 4-1 


CD 


Ph 


3 


cd 


H 


•rl 


CD 


S cd CJ 




CD 


Cd O rH 


CD 


P^ 


•H 


44 




CJ 


CJ 


rH 1H CO 


MH MH 


pel 


s-i o cd 


< 


U-4 44 


e 


cd 


cd 


C 




rH O 


O O 




o x) u 




O -H 


cd 


X) 


4-1 


cd 


rH 


O 4J rH 




rH 


4J 4J CD 


3 


3 


X 




cd 


s 


cd 


ij o cd 


• ■ 


cd 


CJ CD X) 


cd 


• CJ 


CD 


CD 


Q 


•H 


3 


3 O 3 


o 


3 


O O CD 


CJ 


o cd 




43 




Pn 


O 


W Q O 


3 3 


o 


P Pu Ph 


■H 


3 Ph 


X) 


H 


MHCd 


1 


•H 


•H 




•rt 




Sh 




3 




O CD 


s 


4-1 


4-1 




4-1 




CD 




cd 




CD 


o 


cd 


• • cd 


• • 


cd 


> • • 


i 


• • 




• 


CD -H 


3 


S3 


H cm 3 


rH CM 


3 


rH CN CO 


<c 


rH CM 


X) 


CD 


a) 44 














CD 


CD 


a -H 


• 












rH 


CD 


SH CD 


1— 1 


• 


• 




a 


• 


43 


>> 


3 U 


M 


< 


pa 




U 


Q 


6 


T-t 


O CD 














CD 


cd 


CO > 














co 


3 


•H 














CO 


cd 


"2 s 














cd 




5 p 
















CD 


cd 






^^ 








3 


43 


CD £ 






/-n 43 








0) 


44 


CD ° 






r-l a 








<D 




ft 4 - 1 






Cd S-l 4. 


J 




/— 




43 


3 


^ CD 






!-i cd f 






g 






•H 


H CD 






CD CD c 


J 




CD 




X) 




cd 






T3 CD i 




^-^ /^ 


4J 




cd 


rH 


<4-l M 




CD 


CD CD C 


I 


cd CO 


CO 




43 


3 


o 




CJ 


MH SH -r 


^ 


4-1 13 


> 






4H 


cd 




•H 




3 


Cd rH CD 


CO 


~ 


CD 


CD 


C 4-1 




4-1 


"T3 C 


r 


XI CD XI 




CO 


CD 


CD 


o cd 




CD 


H (1) [i 


3 


•H CO rH 


CJ 


B 


a 


3 


•H Q 




•H 


cd u 




CD CD 14-1 x) CD 


<; 


CD 


S-i 




44 




4J 


CO U O « 


3 


CJ 4-1 rH -H 


Ph 


•H 


3 


44 


cd w 




cd 


CD ^n CD CD 




3 cd >% CD MH 


g 


3 


o 


O • 


CJ U 




4-1 


3 rH X) 3 CI 


3 


CD &0,Q -H 


cd 


CO 


3 CD 


•H <| 




CO 


3 cd CD O t 


iO 


•H CD <4H >^ 




43 44 




44 


MH 






CD U MH a. C 




CJ Sh Cd ,Q 




CJ CO 


CD 


U r-\ 


■H CD 




3 


> CD 3 CD T 


H 


CO 60 4-J 43 >-, 


J^ 


CD O 


3 


O 3 


CD 43 




O 


CD X) O X 


3 


bO cd CJ 42 CD 


4-> 


g CJ 


o 


CD 


CD 4J 




•H 


Pi CD 3 « r- 


H 3 


S-i cd XI S-i XI 


rH 




•H 


CD CD 


cd 




4-1 


MH ^— ' rH T 


H 


O ^w cd CD 3 


cd 


00 44 


S-i 


^-t S-i 


rH 3 




cd 


rH cd : 


3 -H 


LH CD XI 3 


CD 


3 CJ 


cd 


42 


U -rl 




CJ 


Cd «£ 4J PC 


4 4J 


0<f «) filH 


33 


•rl CD 


> 


cd MH 






3 


S-l rH CJ O 


cd 


0) N N HI 3 




X) SJ 




CD O 






X) 


CD cd S-i 4-> S- 


J X3 


3 C^ CT\ Bi MH 43 


MH 


3 -H 


CD 


CO 






W 


3 S-i cd ^ C 


> 3 


O rH rH CJ 


O 


3 XI 


CD 


3 3 


• 


CD 




01 CD CD M- 


4 3 


■H S-i 43 S-i 




Pn 3 


HJ 


O 


CO 




a 


Sh 


O X) CO 44 


O 


4-1 |>4 >H o o cd 


CD 


P M 


43 


44 -H 






•H 


O 


CD CD CD 


) Ph 


Cd &H pi, M-l U CD 


CD 




44 


O CO 


CD 




44 


UH 


■^J MH S-l o c 




w). cd co 


4-1 


XI X) 




3 co 


H 




CD 




3 CJ C 


) CD 


■H O O CD CD CD 


3 


3 3 


e 


3 


43 




•rl 


>-l 


rH O X) V 


1 O 


rH 4J 4J CD CO Sh 


4-1 


cd cd 


o 


u o 


cd 




4-1 


CD 


cd 3 CD 4J 4. 


i 3 


42 S-l CD 


■H 




SJ 


CD CO 


H 




cd 


4-1 


3 ^— ' S-I CJ -r 


1 CD 


O ~a- rH 3 Sh rH 


4-1 


CD 44 


C4H 


s: -h 






4J 


3 


O O CD X 


J -H 


>£> r~~ 4-i cd 


CO 


XI CJ 




44 XI 




co 


CD 


•r-f ^H CD s-i x 


1 CJ 


H cn o> H H S-i 


3 


r-\ CD 


cd 


•H 






CJ 


4-1 Cd 3 -H < 


: co 


Cd rH rH XI Cd CD 


M 


CD Sh 


44 


CD CD 




rH 




cd 4-1 O X) 




S-l 3 4-1 XI 




•H -H 


cd 


43 




cd 


^H 


a o cu 3 t 


) <-t 


CD >-i >h CD O CD 


rH 


Ph O 


X) 


CD 44 




•rl 


cd 


3 H CO M q 


! cd 


XI Ci-i Ph p,H [h 


cd 






43 







3 


xl = 


I 3 


CD X 


3 


>-, >> 


CD 


MH 




s 


o 


W pi 


O 


fn W 


o 


cq M 


43 


o o 




cd 


•H 




•H 




■H 




4-1 


44 




PS 


4-> 




4-1 




4-1 






3 




•H 


cd 


• 


cd 


. . 


cd 


• 


S-l 


XI O 




Pk 


3 


rH Cn 


3 


rH CNl 


3 


rH CM 


CD 


CD -H 
















44 


00 44 
















MH 


XI CJ 




• 


. 




, 




. 


< 


3 CD 






i— i 


< 




m 






CJ 




cd 


•r-l CO 



-10- 



It is interesting to note that only three sets of data spanned the entire 11- 
year period: the NSF-CASE series (total federal obligations for academic institu- 
tions); the NCES series on earned doctoral degrees by disciplines; AAUP series 
on faculty salaries and size. In the last case, there has never been a break- 
down of the data by academic disciplines, while such a breakdown wasn't intro- 
duced into the NSF-CASE series until FY 1971. 
Analytical Methods 

The general procedure followed in treating the data may be described as 
"trend analysis," which involved two modes of comparisons of changes over time 
in measures of educational variables: (a) differences among various types of 
institutions in the trend for a given variable; (b) concurrent trends among 
two or more variables for a given type of institution. Combinations of both 
modes of trend analysis were also employed within a single set of comparisons. 

Three types of statistical measures or indices of changes in the financial 

and educational variables were used: 

The arithmetic mean . With some reservations, this measure of central 
tendency for the distributions of all of the variables was selected in 
preference to the median. Despite being subject to undue influence by 
extreme scores in skewed distributions, means and their accompanying 
standard deviations posed simpler computational problems and reflected 
more precisely the actual magnitudes and dispersions of the institutional 
"scores" on the several variables. 

In computing the means for the successive years in a trend series, it was 
decided to use the same number of cases throughout — even though this meant 
the discarding of data for substantial numbers of institutions. For 
example, in the case of NCES' HEGIS data, only 100 of the 148 institutions 
in the sample had complete records for all of the financial variables 
used. This was the greatest reduction required for any variable used in 
the analysis, however, and the 100 universities included 93 (64 per cent) 
of the 145 institutions comprising the population of the top three Carne- 
gie Commission categories. It is believed that population proportions 
as high as this (or higher) have yielded means sufficiently reliable for 
the kinds of trend comparisons undertaken in this study. 



-11- 



In order to check directly on the kinds of variations that might result 
from using different numbers of institutions with different variables, the 
principal analyses were duplicated with data from the residue of 63 insti- 
tutions (34 private and 29 public) whose records were complete for all 
variables and relevant years. Unfortunately, this reduced sample of 63 
universities was quite unrepresentative of the original 148, since it 
contained 62 per cent of the private but only 31 per cent of the 
public institutions shown for the total sample in Table 2. Nevertheless, 
the trend patterns for the means for "All Institutions Combined," "All 
Private Institutions" and "All Public Institutions" showed reasonably 
good agreement with those derived from the larger samples. Further break- 
downs into sub-classes of institutions produced such small numbers of cases 
that the variability increased considerably. 

Index numbers . For purposes of comparing trends in the means for different 
groups or for variables with different units, it seemed desirable to have 
a "common-denominator" to which the various trend series could be reduced. 
For this purpose, an index-number series was created for each variable by 
selecting a base year and expressing the means for all other years in the 
series as percentages of the mean for the base year (multiplied by 100) . 
The year 1971-72 was chosen as the base year for all trend series, and 
its mean in each case was assigned the index value "100." 

Percentage comparisons . Several of the questions posed by the Panel 
called for the calculation of percentage relationships between a given 
variable and a reference variable. For example: "What proportion of the 
educational and general revenues of universities is provided from federal 
R&D funds?" Such relationships expressed as percentages may be assumed 
to be generally comparable from year to year in a series, for the same 
group of institutions. 

Price Indices 

In attempting to assess the impact of federal R&D funding upon universities, 

it seemed highly important to take into account the effects of inflation upon 

the financial trends under study. For this purpose, four series of price indices 

have been used as deflators — each appropriate to a given type of expenditure: 

3 

1. Halstead's Higher Education Price Index 

2. Halstead's Construction Price Index"* 

3. Halstead's Equipment Price Index^ 

4. The NIH R&D Price Index 4 



D- Kent Halstead, Higher Education Prices and Price Indexes . Washington, D.C. t 
U.S. Government Printing Office, 1975. 



The NIH R&D deflator was recently developed by Westat, Inc. under contract with 
the National Institutes of Health. 



-12- 



Price-index series have been compiled for all four of these deflators — 
using the FY 1964 as the base year. They are presented in Table 4 below, 
together with parallel series for the Consumer Price Index and the Wholesale 
Price Index. The latter two series have been derived from data published in 
the Economic Report of the President, 1975 and in Economic Indicators published 
by the Council of Economic Advisors. Since both federal series are stated on 
a calendar-year basis (except for the current year, when monthly or quarterly 
figures are given) , they have been converted to a fiscal-year basis by averaging 
values for continguous calendar years. 



Table 4 . Price Deflators for Four Types of Higher-Education 

Expenditures, together with the Consumer Price Index 
and the Wholesale Price Index 



Fiscal Halstead Halstead Halstead 

Year Higher Ed. Construction Equipment 

Price Price Price 

Index Index Index 



NIH 


Consumer 


Wholesale 


R&D 


Price 


Price 


Price 


Index 


Index 


Index 







1964 



100.0 



100.0 



100.0 



100.0 



100.0 



100.0 



1965 


104.3 


103.0 


100.2 


102.6 


101.5 


101.2 


1966 


109.5 


106.8 


101.3 


106.5 


103.9 


103.8 


1967 


115.2 


111.9 


105.2 


111.2 


106.8 


105.6 


1968 


122.1 


120.0 


108.4 


117.6 


110.6 


107.1 


1969 


130.4 


129.2 


111.3 


124.0 


115.9 


110.5 


1970 


139.4 


138.7 


116.1 


130.9 


122.5 


114.7 


1971 


148.3 


150.7 


121.5 


138.5 


128.7 


118.6 


1972 


156.5 


163.0 


124.0 


144.3 


133.6 


123.2 


1973 


164.5 


173.2 


127.8 


150.5 


140.0 


134.1 


1974 


176.0 


184.9 


137.8 


160.2 


152.1 


155.8 



-13- 



III. TRENDS IN TOTAL EDUCATIONAL-AND-GENERAL REVENUES 

The term "educational-and-general revenues" designates the total income 
of an academic institution from all sources for its regular educational opera- 
tions and their support functions. These and other financial data are collected 
annually by the National Center for Education Statistics (NCES) in one of its 
Higher Education General Information Surveys (HEGIS) . 

Excluded from the category of "Educational-and-General Revenues" in the 
HEGIS financial survey are funds for such purposes as student aid, auxiliary 
enterprises, and "major service programs" (e.g., hospital operations or Federal- 
ly Funded Research and Development Centers) . 

Trends in E&G Revenues in Constant Dollars 

The general paradigm for presenting most of the financial results of this 
study is illustrated in Tables 5 and 5A on the following page. Mean E&G reve- 
nues by type of institution in thousands of constant dollars are shown in 
Table 5 — for the six fiscal years 1969 through 1974. It was decided as a 
general rule to omit parallel tables showing means in current dollars, partly 
to save space, but mainly to focus attention on the financial condition of the 
institutions, which is best represented by trends in "real-dollar" equivalents. 
Moreover, the current-dollar mean values may be derived from the constant- 
dollar means through multiplication by the appropriate price-index values shown 
in Table 4. 



Prior to FY 1969, NCES included funds for Federally Funded Research and Develop- 
ment Centers in the E&G category under "sponsored research". This meant that 
the E&G revenue data supplied by NCES for FY 1966 through 1968 could not be used 
in the present study because of incompatibility with later data. 



-14- 



ON 



CO 

on 



On 



ON 
MO 

on 



MO 


p*- 


r^ 


en 


on 


CO 


r-~ 


oo co 


rH 


CM 


o 


<r 


CM 


— 1 


r^ 


CO o 


oo 


M0 


ro 


rH 


00 


in 


o 


CM CO 


















oo 


r~ 


O 


rH 


ON 


<r 


oo 


cm m 


en 


en 


in 


CM 


CO 


m 


CM 


in cm 


co- 
















en 


o 


ro 


O 


00 


CO 


oo 


ON C~ 


en 


H 


O 


<t 


l>» 


CM 


r-t 


ON M0 


en 


^3- 


O 


o 


CO 


CO 


<f 


m cm 


r> 


r*> 


o 


iH 


00 


CO 


MO 


rH -J 


ro 


m 


m 


CM 


CO 


m 


CM 


m cm 


</> 
















CM 


-3 


-cr 


m 


ON 


ON 


CM 


-3- o 


<r 


iH 


r^ 


CM 


on 


MD 


oo 


M0 CM 


ON 


00 


in 


CM 


rH 


ON 


CO 


f-. r-H 


















r> 


00 


H 


CM 


r^ 


rH 


m 


•h -a- 


CO 


CO 


m 


CM 


ro 


m 


CM 


m cm 


u> 
















U1 


00 


r~ 


r~ 


r*. 


<on 


-3 


m m 


oo 


00 


m 


<T 


CM 


MO 


on 


■-I <n 


o 


en 


On 


CO 


CO 


CO 


rH 


rH O 


















r~ 


r- 


00 


CM 


M0 


rH 


m 


o -J 


co 


co 


<r 


CM 


CO 


m 


CM 


m c i 


co- 
















cm 


00 


r-~ 


MD 


rH 


<r 


MD 


-3- o 


rH 


m 


O 


O 


On 


co 


ON 


CM o 


en 


rH 


in 


<r 


m 


oo 


CO 


M0 O 


■1 


w 


*. 


M 


n 


•1 


•■ 


»• •> 


M0 


r~ 


00 


CM 


m 


O 


m 


ON CO 


CO 


en 


<r 


CM 


CO 


m 


CM 


~3 CM 


CO- 
















r^ 


c-~ 


m 


ON 


r-. 


in 


CO 


00 M0 


-J- 


o 


en 


CO 


r- 


on 


^D 


MS CM 


ON 


ON 


On 


u-i 


c-t 


r-~ 


vO 


00 o 


















•tf 


M0 


r~. 


CM 


CO 


r~- 


rH 


r-~ CM 


CO 


en 


-3 


CM 


CO 


<r 


CM 


-a- cm 


CO- 

















01 

3 
Z 



o 


M0 MO O 


-3 <■ O 


o o 


o 


-J CM CM 


m CM CO 


m m 



c 
o 

«4H -H 
O 4J 

3 
HI 4-1 
D.-H 
S-> 4J 
H CD 
C 









m 






tn 







tn 




H 






rH 


[0 O 




C 




O 


m 




o 


rH JS 




o 


en 


o 


C 


tn 


o 


o o 




•H 


rH 


x: 


o 


rH 


-C 


O 00 




u 





o 


•H 


O 


o 


rC 




3 


o 


00 


4-1 


o 


00 


O rH 




4-1 


-C 




3 


-C 




co n) 




tH 


u 


rH 


4-1 


(J 


rH 


o 


to 


iJ 


00 


n) 


•H 


00 


tfl 


rH T-l 


c 


CO 




o 


4-1 




'-J 


« OO 


o 


C 


rH 


•H 


tn 


rH 


•H 


O 01 


■rf 


l-l 


CO 


-o 


a 


ra 


13 


■H S 


4-1 




CJ 


0) 


H 


o 


01 


-O 


3 


cu 


•H 


g 




■H 


S 


0) 4-1 


4-1 


u 


X< 




o 


•3 




S 3 


H 


CO 


0) 


4-1 


•H 


01 


4J 


o 


4-1 


> 


s 


3 


^H 


s 


3 


J3 rC 


0) 


•H 




o 


_C 




O 


U 4-1 


C 


-J 


Xi 


-C 


3 


j: 


43 


•H -H 


KH 


Oh 


4J 


4-1 


Cm 


4J 


4J 


S 3 






•H 


■H 




■H 


•H 




rH 


rH 


3 


3 


rH 


3 


3 


rH rH 


< 


rH 

< 






5 






^^ 



s 
o 

•H 

4-1 

CO 

o 
3 
•O 

1-4 
O 



u 

0) 

4J 

c 

01 
CJ 



3 

o 

00 



o 

to 



0) 
3 

a 

0) 

> 

0) 
OS 



CO 

1-1 

0) 
£3 
4) 
O 
I 

•c 

c 

CO 
I 

rH 
CO 
C 

o 



CO 

o 
3 
•3 

W 



U) 




c 




CO 




0) 




E 


> — - 




o 


0) 


o 


j= 


rH 


4J 






II 


(-1 







CM 


U-i 


r-» 




o> 


tn 


rH 


u 




0) 

J2 


£ 


B 




3 


u 


Si 


o 




14-1 


X 




CD 


tn 


T) 


c 


c 


03 


H 


d) 




s 


c 


S — r 


■H 






m 


til 




T) 


01 


c 


rH 


01 


rQ 


r4 


tn 


t-l 


H 



4-> 


<! 




O U 


CO 


m 




3 


z 






CU 4-1 




0) 




0.1H 




rH 




r^ W 


*• 


rQ 




H (0 


01 


nf 




c 


CJ 


H 




M 



u. 





rM 


u 


r*. 


CO 


CT\ 


CU 


rH 


!H 




rH 




CO 




O 




tn 


rH 



u 

01 

XI 
B 
3 
Z 



mo r^ in 

ON 0> Ci 



r— -J- o 

O O rH 



rH -3- 

o o 



a 


* ISM- 


CO CM -d- 


OO o 


o 


Q\ 0\ <Ti 


O O O 


o> o 



o 
o 



o o o 
o o o 



o o o 
o o o 



o o 
o o 



mo <r o 

O-v ON O 



ON 00 ON 

ON ON ON 



VO ON 
ON ON 



m 

ON 



CN4 

ON 



in ~3 o 
on on o 



m r~ cm 

ON ON ON 



m m 

ON ON 



m CM rH 

ON on o 



on cm in 

00 ON 00 



CM rH 
ON ON 



a 


lOiOO 


-a- <r o 


o o 


o 


^3* CM CM 


m cm co 


m m 









(0 






CO 


o 




tn 




rH 






rH 


CO o 




C 




O 


to 




o 


rH rC 




o 


tn 


o 


B 


tn 


o 


o a 




■H 


rH 


-C 


o 


rH 


rC 


o 00 




4-1 


o 


CJ 


•H 





o 


J3 




3 


o 


CO 


4-1 


o 


CO 


CJ rH 




4-1 


rC 




3 


rC 




CO CO 




•H 


CJ 


rH 


4-1 


u 


rH 


CJ 


CD 


4-1 


CO 


CO 


•H 


CO 


CO 


rH -H 


C 


CO 




CJ 


4J 




U 


CO -O 





c 


H 


•H 


tn 


rH 


•H 


O 01 


H 


hH 


CO 


T3 


13 


CO 


-a 


■H S 


4J 




U 


01 


M 


u 


a) 


•o 


3 


0) 


■H 


S 




■H 


s 


0) 4J 


4-1 


4-1 


X) 




o 


•o 




S 2 


H 


CO 


0) 


4-1 


•H 


0) 


4-1 


o 


4J 


> 


S 


3 


rH 


S 


3 


rC J3 


tn 


•H 




O 


XI 







4J 4-1 


c 


l-i 


^ 


XI 


3 


X 


-C 


•H -H 


H 


Ch 


4-1 


4J 


PL, 


4-1 


4-1 


3 3 






•H 


•H 




■H 


•H 




r^ 


rH 


3 


3 


rH 


3 


3 


rH rH 


— 1 


rH 






rH 






rH rH 


< 


< 






< 






< < 



-15- 



The E&G means for all of the 100 institutions in the sample which supplied 
data for all six years of the series are presented in the first line of Table 5. 
These figures show a slow, progressive increase in average E&G revenues from 
$34.9 million in FY 1969 to $38.8 million in FY 1974 (an increase of 11.1 per 
cent in constant dollars) . 

The nature of this trend is perceived more readily through inspection of 
the index numbers shown in Table 5A, where the means for other years are expressed 
as percentages of the mean for FY 1972. For all institutions combined, the 
index-number trend increases from 92.1 in FY 1969 to 102.3 in FY 1974 — a rate 
of growth in revenues that will be shown later to have been somewhat lower 
than the overall increase in enrollment. 

Private vs. Public Universities 

It is immediately apparent from the data in Tables 5 and 5A that the 
E&G revenue trends for private and public institutions are quite different. 
The latter show a positive growth trend throughout the six-year period, with 
appreciable acceleration from FY 19 72 to 19 74. The private institutions, on 
the other hand, show a slower rate of revenue growth through FY 1972, and then 
a significant drop from the index value of 100.0 for the latter year to 96.4 
and 96.9 for FY 1973 and 1974, respectively. 

Presence or Absence of Medical Schools 

When private and public institutions are combined on the basis of medical- 
school status, all institutions without medical schools show a somewhat more 
rapid rate of growth than those with them. However, when this type of break- 
down is made within the private and the public sectors separately, different 



-16- 



patterns emerge. Private institutions with medical schools increased some 5 

per cent in E&G revenues while those without them decreased 6.3 per cent from 
FY 1969 to 1974. 

Both types of public institutions had progressive E&G revenue increases 
throughout that period, but those without medical schools had a much greater 
overall growth rate (30 per cent vs. 14 per cent for the group with medical 
schools) . 
Trend Differences among Carnegie-Commission Categories of Institutions 

Since the Carnegie Commission's classification was established mainly 
in terms of doctorates and research funding, comparisons were made among these 

categories in terms of E&G revenue trends. The results are shown in Tables 6 
and 6A on the following page. 

The most important finding is that the constant-dollar decline in revenues 
noted in Table 5 for the private institutions after FY 1972 is attributable 
entirely to Research Universities I (the top-ranking category) . Private Research 
Universities II and "Other Categories" showed positive growth during this period, 
with the latter group having the higher rate of increase (6.5% vs. 2.5%). From 
FY 1969 to 1974, the overall percentage changes were: Research Universities I, 
-2.5%; Research Universities II, 4.1%; Other Categories, 21.1%. 

Among the public institutions, "Other Categories" likewise showed the 
highest rate of overall E&G revenue growth (29.1 per cent). For the two groups 
of research universities, the increases from FY 1969 to 1974 were: Research 
Universities I, 17.9%; Research Universities II, 15.6%. 



-17- 







1^ 0100 VO 


CTl tO OO CTi 


















■* 


cm 10 *a- to 


CM r- tH to 




in 




«* 


oi^mi/i 


rH rH r-,H 




r-~ 


O CM CTl CM 


OO CM rH iH 




(D 




r-~ 








CTl 








3 




CTi 


vDMN lO 


t-^ vO "H- rM 




i— i 


t-. rH !-H O 


CTi CTl r- CM 




3 




<—t 


CTi CTl O O 


O O O rH 






IO vO N N 


tONtON 




a> 






rH rH 


rH rH rH rH 






^5- 






> 

ID 

OS 














OCO^N 


OO t^ CD LO 
















to 


i— 1 *t K) CTi 


NOION 




rH 




to 


>=3- O «* CM 


CM CM r~- CM 




r^- 


*3" CM CTl LO 


to iH o r^ 




CO 




r^ 








CTl 








rH 




CTi 


vO to CM IO 


to tO rH LO 




i— i 


t~- H H Ol 


CONlOO 




CD 




<-t 


CTl CTl O O 


O CD O O 






K)\ONH 


to r— to cm 




3 






rH H 


rH rH rH rH 






■foO- 






0) 
CJ 














^- U1CO 00 


CTl O CTl CD 




T3 












rsi 


i— 1 «* r-H CTl 


CTl to to LO 




3 




CM 


O O CD O 


O CD O CD 


u 


r~- 


Of *o> 


>h r^ «a- r-~ 




CO 


rH 


r- 






03 


cti 








i 


Cfl 


CTi 


O O O O 


O CD O O 


CU 


1— 1 


oo i-n i— t oo 


t~^ «n- lo cn 




rH 


CD 


rH 


O O O CD 


O O O O 


>h 




MvONH 


tONIOH 




CO 

3 


>H 




rH rH rH rH 


rH rH rH rH 


tH 










o 


r-i 








03 










•H 


Cfl 








U 




co oo *3- -ri- 


t-~- rH CM m- 




4-> 


O 








in 


1— ( 


CO CD vO CTi 


CM O 00 CTl 




CO 


in 


^-i 


to to to r-~ 


OlOHN 


•H 


t~- 


tO O O CTi 


COH NH 




(J 


■H 


r^ 






Hh 


CTl 








3 


Ph 


CTi 


\OiO CO ^" 


CTi o co r-~ 




i— i 


MvONH 
CO LO tO U"> 


vO LO **■ CTi 
tO t~- tO rH 

rH O tO CM 




of Ed 
100) 




rH 


CTi CTi CTi CTl 


CTl CD CTi CTi 
rH 




o 


U1 CM r~~ tO 


CTl "* LO , ^- 




m n 




O 


t~^ LO ^ CM 


I^HrJ-Ol 




t-~ 


rH I— 1 O »H 


lo oo iH >-n 


■ 


C 




r~. 








CTi 






in 


co cm 




CTl 


l-O \D 00 O 


LO \0 vO tO 




r-H 


tO \D CM tH 

■foe- 

t-~ co r-~ co 


LO rH •rj- 00 
tO t-» tO iH 

r~- "* o co 


u 

•H 
HH 

in 

•H 

+-) 


the Me 
FY 197 




rH 


CTl CTi CTl CTl 


CTi CTi CTi CTi 




en 


o lO l-» CTl 


r-» r^ cm ti- 


cfl 






CTi 


rH O -H; CTi 


inoioco 




\0 


CTl 00 O vO 


CM CM rH iH 


4-> 


rH iH 




VO 








CTl 






oo 


O O 




CTi 


UO O 00 t~~ 


CTl O O O 




rH 


tO vO CM rH 

•fo<3- 


tO vO tO rH 


c 
o 


<-H 4H 

in in 




<—< 


CTl CTi CTi CO 


OO CTl CTi CO 








•H 


U 3 














4-> 


CD CO 














CO 


43 CD 








H 






o 


E2 


rH 






O 






3 


3^ 


CD 






43 


O CTl CM LO 


«3" CM CO ** 


T3 


2 


43 


\DOlN in 


^NOO* 


E 


«* rH rH iH 


inHHN 


W 


vD 


s 


•cj- rH rH rH 


LO rH rH CM 


3 








X 


3 






2 






Si 

o 


CD CD 
T3 rH 

3 42 


2 








hH 


hH 




H CO 




KH 


t—i 




t— f HH 


H H 


rH 


E- 




1 — 1 hH 


rH t-H 




t/) 




CD 


C 




in 






C»l« 


in in (/i 


4-> 


■H C 




3 m in 


in in in 




O 0) 0) 


C CD CD 


c 


•H 




O CD CD 


3 CD CD 




• H -H -H 


O -H -H 


CD 


in 




•H -H 'H 


O -H -H 




4-1 4-> 4-> 


•H 4-> 4-1 


C_3 


-3 3 




4-> ■!-» +J 


•H 4-1 4-1 




3 -H -H 


+■> -H 'H 




3 S 




3 -H -H 


4J 'H -H 


3 


4-> w in in 


3 in in in 


rH 


OD O 


3 


+-i in in in 


3 in in in 


o 


•H (h h 0) 


+J rH f-l CD 


CO 


Sh 4= 


O 


•H rH rH CD 


+-> !h in CD 


4H -H 


+J 0) UH 


•H CD CD -H 


3 


E- W 


<HH -H 


4-> CD CD -H 


•H (D CD -rH 


O +-> 


in > > Sh 


4-1 > > rH 


c 




O 4-> 


in > > rH 


4-" > > rH 


3 


3 -H -H O 


in -H -H o 


•H 




3 


C 'H -H O 


m -h >H o 


o> 4-> 


ih C C OO 


3 C 3 60 


4-> 


• 


CD ■(-> 


rH C 3 OO 


3 3 3 60 


Ph-H 


33 33 CD 


HDD 11) 


CO 


< 




(h-H 


33 -D CD 


hH D 33 CD 


>>4-> 


0) 4-> 


4-> 


2 


vO 




x+-> 


CD +-> 


4-> 


f-i in 


■u 42 42 w 


u 45 42 co 








H in 


4-> 42 42 CO 


a 42 42 cfl 


c 


CO u o u 


•H u u u 




CD 




3 


1JUUU 


•H U uu 


1 — 1 


> H H 


rH J-, |_ 




rH 




i — i 


> rH !h 


rH rH Sh 




•H CO CO Sh 


42 CO CO rH 


CD 


43 






•H CO Cfl rH 


43 Cfl CO rH 




Jh CD CD CD 


3 CD CD CD 


U 


CO 






)h CD CD CD 


3 CD CD 




O. m »i£ 


a, m in 42 


rH 


H 






Cl, in in 42 


EL, in in 42 




0) CD 4-> 


CD CD 4-1 


3 






CD CD «-> 


CD CD 4-" 




HKoiO 


HtfPJO 


O 






HQIKO 


rH OS 04 O 




.H 


rH 


C/} 






rH 


t-i 






< 


< 


CO 










< 


< 



-18- 



Comparative Data from Another ACE Study 

Since the final year of the present study's data base was FY 1974, the 
results couldn't reflect the effects of the sharp cost escalation that has 
occurred since that time. It seemed desirable, therefore, to cite results 
from a recent study that extended through FY 19 75, using more refined measures 
of cost inflation. Excerpts from the results of that study are summarized in 
Tables 7 and 7A on the following page, and a footnote to these tables gives 
the reference to the study. 

The data in Tables 7 and 7A — derived from an ACE special survey — are 
based on E&G expenditures per full-time-equivalent (FTE) student for the top 
three Carnegie Commission categories of universities, covering the three fiscal 
years 1973, 1974, and 19 75. The median values in constant dollars are presented 
in Table 7, and inspection of these figures shows both the expenditure levels 
and trends over the three-year period. But the magnitude and direction of the 
trends are measured more precisely by the percentage-change values shown in 
Table 7A. The latter were derived by computing such percentages for individual 
institutions, and then computing the median of these percentages for each 
category of institutions. 

In general, all three groups of private universities had percentage decreases 
in constant dollars per FTE student for both FY 1974 and FY 1975 (the negative 
"growth" increasing for the latter year) . Although not strictly comparable 
with the trend data shown in Table 5A for the present study, the two sets of 
results are consistent for their one common year (FY 1974) . 



-19- 



Table 7: Median Educational-and-General Expenditures per 
Full-Time-Equivalent Student by Carnegie Commis- 
sion Categories (Constant Dollars) 



Type of 
Institution 



Number 



1973 



Fiscal Year 



1974 



1975 



Private Institutions 36 

Research Universities I 12 

Research Universities II 11 

Doctoral Universities I 13 

Public Institutions 60 

Research Universities I 21 

Research Universities II 16 

Doctoral Universities I 23 



$6,510 
3,312 
2,570 



2,956 
2,078 
1,818 



$6,393 
3,357 
2,354 



2,841 
2,262 
1,871 



$6,326 
3,075 
2,248 



2,823 
2,153 
1,735 



Table 7 A. Median Percentage Change in Educational-and- 
General Expenditures per FTE Student by Carne- 
gie Commission Categories (in Current and in 
Constant Dollars) 



Type of 
Institution 



Number Current Dollars 
FY 19 73 Fif 19 74 
to 1974 to 1975 



Constant Dollars 

TT7T 



FY 
to 



1974 



FY 
to 



T571T 
1975 



Private Institutions 

Research Universities I 

Research Universities II 

Doctoral Universities I 

Public Institutions 

Research Universities I 

Research Universities II 

Doctoral Universities I 



36 




12 


4.7% 


11 


1.7 


13 


5.1 


60 




21 


5.9 


16 


7.8 


23 


10.3 



5.8% 

1.5 

4.2 



5.9 
5.9 
8.6 



-2.0% 

-4.8 

-1.7 



■0.9 
0.9 
3.2 



-3.8% 

-7.7 

-5.3 



•3.7 
3.5 
1.3 



The data in Tables 7 and 7A were published in a monograph by 
Lyle H. Lanier and Charles J. Andersen entitled A Study of the 
Financial Condition of Colleges and Universities: 1972-75 
(ACE Special Report, October ,1975) . 



-20- 



In the case of the public institutions, the trend data in Table 5A all 

show greater positive growth for FY 1974 than those in Table 7A. But the 

differences might be due largely to the fact that in the earlier study median 

expenditure per FTE student in constant dollars was the average used, whereas 

in the present study the means of actual constant-dollars revenues were used 

(hence not taking into account enrollment changes) . 

It seems reasonable to conclude — in the light of inflation-recession trends 

in the national economy — that if the present study had been able to include 

E&G trend data for FY 19 75, a progressively worsening financial picture for 

the research universities would have emerged. This conclusion is supported by 

6 



a recent article by William G. Bowen. 



Bowen, William G. The Effects of Inflation/Recession on Higher Education. 
Educational Record , Summer 1975, Volume 56, No. 3. 



-21- 



IV. TRENDS IN REVENUES FOR SPONSORED RESEARCH 

Since the basic-research capability of the nation resides largely in the 
research universities, it seems obviously important to know the nature and 
extent of changes that might have been occurring in the financial support for 
university-based research in recent years. This problem has been studied through 
the examination of three sets of trends: 

1. Trends in the proportions of R&D funds in the total educational (E&G) 
budgets of universities in the sample. 

2. Trends in the total support for university R&D in all fields, by 
types of institutions, in constant dollars. 

3. Trends in federally sponsored R&D in all fields by types of institu- 
tions, in constant dollars. 

R&D Proportions of Total E&G Revenues 

Research and development at universities are supported by funds from 
various sources but by far the greater part comes from federal agencies. The 
institutions report annually to the National Center for Education Statistics 
the total amounts provided for sponsored research, with breakdowns by major 
sources. Hence, percentage relationships between these R&D revenue components 
and total E&G revenues may be computed. 

The present analysis has been limited to calculations of two sets of per- 
centages for various groupings of the sample of institutions: (a) total R&D 
revenues as a percentage of total E&G revenues; (b) federal R&D revenues as a 
percentage of total E&G revenues. The difference between these two types of 
percentage values obviously gives directly the proportion of non-federal funding 
of R&D activities in the E&G budget. (To conserve space, these residual 



-22- 



percentages have been omitted from the summary of the results in Table 3 on 

the following page.) The trends in the two sets of percentages are shown for 

the six fiscal years (1969-1974) for which NCES data were available — first for 

all institutions combined and then for the various sub-groups within the total 

sample shown in earlier tables: 

Trends for all institutions . The first line of Table 8 shows that total 
R&D revenues decline slowly from 21 per cent of E&G funds in FY 1969 to 
19 per cent in FY 1974 (the decline in percentage points represents a 
drop of 9.5 per cent). 

The corresponding decline for federal R&D funding, shown in the second 
line of the table, is somewhat greater: a drop from 18 per cent of total 
E&G revenues in FY 1969 to 15 per cent in FY 1974 (representing a percen- 
tage decline of 16.7 per cent). 

Another method of comparing the two R&D variables is to calculate the 
percentage relationship between the federal R&D and the total R&D compon- 
ents. For example, in FY 1969, federal R&D funds accounted for 85.7 per 
cent of total R&D support; but by FY 1974 this proportion had declined 
to 78.9 per cent. 

All private institutions . The second section of Table 8 shows similar 
percentage data for all private institutions; and within the private sec- 
tor, corresponding percentages are shown for institutions with and those 
without medical schools. These data show generally that the private 
institutions have higher proportions of R&D funds in their total educa- 
tional budgets than all institutions combined. The trend patterns, however, 
are generally similar to those just described for all institutions com- 
bined: moderate declines both in total and in federally sponsored R&D 
funds as proportions of total E&G revenues. For example, total R&D funds 
for FY 1969 provided 28 per cent of the E&G budget, but the proportion 
dropped to 25 per cent by FY 1974 (a 10.7 per cent decline in percentage 
points.) The decline is somewhat greater for federal R&D revenues: from 
24 per cent of E&G funds in FY 1969 to 21 per cent in 1974 (a 12.5 per 
cent drop) . 

In terms of the relationship of federal R&D funds to total R&D revenues, 
for private institutions the percentage relation for FY 1969 is 85.7 per 
cent, and this value declines only slightly to 84 per cent fy FY 1974. 



-23- 



Table 8 . Trends in Total and in Federally Sponsored R&D Revenues 
as Percentages of Educationalr-and-General Revenues 
(Based on Constant Dollars) 



Type of 
Institution 



Number 



Fiscal Year 



1969 1970 1971 1972 1973 1974 



All Institutions 

Total R&D/Total E&G Rev. 10Q 

Federal R&D/Total E&G Rev. 100 

All Private Institutions 

Total R&D/Total E&G Rev. 46 

Federal R&D/Total E&G Rev. 46 
With Medical Schools 

Total R&D/Total E&G 26 

Federal R&D/Total E&G 26 
Without Medical Schools 

Total R&D/Total E&G 20 

Federal R&D/Total E&G 20 

All Public Institutions 

Total R&D/Total E&G Rev. 54 

Federal R&D/Total E&G Rev. 54 
With Medical Schools 

Total R&D/Total E&G 24 

Federal R&D/Total E&G 24 
Without Medical Schools 

Total R&D/Total E&G 30 

Federal R&D/Total E&G 30 

All with Medical Schools 

Total R&D/Total E&G Rev. 50 

Federal R&D/Total E&G Rev. 50 

All without Medical Schools 

Total R&D/Total E&G Rev. 50 

Federal R&D/Total E&G Rev. 50 



21% 


20% 


19% 


20% 


20% 


19% 


18 


17 


16 


16 


16 


15 


28 


27 


26 


27 


27 


25 


24 


23 


21 


22 


21 


21 


29 


29 


28 


28 


29 


27 


25 


24 


22 


23 


23 


22 


26 


25 


24 


25 


24 


23 


23 


22 


20 


21 


20 


19 


16 


14 


13 


14 


14 


13 


13 


12 


11 


11 


11 


10 


17 


17 


16 


16 


17 


16 


14 


14 


13 


13 


13 


13 


14 


12 


11 


12 


12 


11 


12 


10 


9 


9 


10 


9 


24 


23 


22 


22 


23 


22 


20 


19 


18 


18 


18 


17 


19 


17 


16 


17 


17 


16 


16 


15 


13 


14 


14 


13 



Source: National Center for Education Statistics, 



See Table 4 for the price indices used as deflators: (a) for E&G 
revenues, Halstead's Higher Education Price Index; (b) for R&D revenues, 
the R&D price index recently developed by Westat 5 Inc. for NIH. 



-24- 



All public institutions . Public institutions generally have lower pro- 
portions of R&D funds in their total E&G budgets than private institu- 
tions. This is due partly to the fact that public universities engage 
in a much wider variety of educational functions than do private institu- 
tions (e.g., substantially larger extension and public service programs, 
as well as many educational programs not conducted by private institutions 
such as those in agriculture and other vocationally oriented curricula) . 

Public institutions show declines similar to those of private institutions, 
both in total R&D and in federal R&D funds as proportions of their total 
educational budgets, but the overall percentage decreases tend to be 
somewhat higher than those found for all private institutions. For example, 
total R&D funds declined from 16 to 13 per cent of total E&G funds over 
the six year period ( a drop of 18.8 per cent); while the decline for 
federal R&D funds over the same period is from 13 to 10 per cent (a drop 
of 23 per cent) . 

Presence or absence of medical schools . The last two sections of Table 8 
show R&D/E&G percentages for institutions with and those without medical 
schools (combining data for private and public universities) . The dif- 
ferences between the two groups of institutions are not great, but there 
is a slight tendency for universities with medical schools to show some- 
what smaller percentage declines over the six year period, for both total 
and federal R&D funds, than is the case for institutions without medical 
schools. 



-25- 



Trends in All Sponsored R&D Revenues 

Although federally sponsored R&D funds constitute most of the support for 
research in universities, it was decided to make separate trend analyses for 
the total amount and for the federal component. The procedure is the same as 
that followed in chapter III for total E&G revenues. Mean values per institu- 
tion in constant dollars are shown in a table for the several years in the data 
series by several groupings of institutions. In an accompanying table on the 
same page, trends in index numbers for the constant-dollar means are presented 
for purposes of ready comparisons of trends. Following this format, the trend 
data for sponsored R&D revenues are shown on the following page in Tables 9 
and 9A, by type of institution. The results for the three main groupings may 
be briefly summarized as follows: 

All institution? combined . Inspection of the index-number values in Table 
9A shows a somewhat more irregular trend for R&D funds than was found 
earlier for total E&G revenues. There is a slow decline from FY 1969 
through FY 1971; then an increase for FY 1972 which was maintained vir- 
tually constant through FY 1973 — followed by a substantial drop in FY 
1974. The decline over the six-year period was approximately 6 per cent. 

Private vs. public institutions . The trend indices in Table 9A show that 
the private institutions follow the pattern just described for all insti- 
tutions through FY 1972, followed by a rather marked decline from the 
index value of 100 for that year to 94.3 for FY 1973 and 91.2 for FY 1974. 

Public institutions, on the other hand, maintained an essentially constant 
level of R&D expenditure from FY 1969 through FY 1972, but then increased 
sharply from the index number of 100 to 109.4 for FY 1973 — followed by a 
drop to 103.8 in FY 1974. Thus, although differing somewhat both in 
pattern and magnitude, the changes after FY 1972 are essentailly similar 
for R&D funding to what was found for total E&G funding: namely, the 
private institutions experienced a relative decrease while public insti- 
tutions had an increase in R&D funding after FY 1972. 

Over the entire period (FY 1969-1974) private universities had a decline 
of 12 per cent in R&D revenues, while public universities showed a slight 
increase of 4.8 per cent. 



-26- 






r-- 

at 





r-j 


h 


r-« 


rt 


01 


01 


i— i 


>h 




rH 




CO 




o 




in 


T-l 



o 

CI 



vO 



rH 

CD 
X 

6 
3 
2 



o 

<+H -H 
O P 

3 
0> P 
P.-H 
>>P 

E-i w 
C 



CT> HOll^ O0000 CNIO 

r» oooo to o co ■<3-i-i 

to o> lo rH cnio>c*i MTt 

00 O *=T \£> O CTi to r-l»3- 

rH rH rH 

oo «3-oct> r~- ** o cni ** 

(vj O O f-~ >DvDI~- C-J to 

t-~ Ni- 1 <NI ini/llO C% LO 

#* n «t * a n m t\ a. 

00 Hi/liO O O to rq ^. 

I— t i— I rH i— I 
■fcr> 

\0 00 00 CT> HlOlfl CN1 rH 

to "* O to O CTi cn] O rH 

t-~ OlrHOO O LO rH «* O 

CO HWN (OOlM N W 

• — I . — t .-I 

in aiLOrsi lo cm *ct r^ to 

LO -tooo O to O oo 

•<* ** «* lo oiuio rH co 

oo i-H »* t — ■ lo en to r-4 ■* 

i-i i— i i— i 

C--1 OlMH CO»ON WH 

r~- NN^ OlMO voce 

OO i— ICSJO O 00 i— I vOO 

O0 CNJ LO 00 vO Ol tO CN1 LO 

I— I ■— t I— I 



O rH to CT> 0)00 «3-0 

CM rH rH tO >*H ID COU1 

CTi ^M vD CTi -3- rH *3" to 

oo csiLOoo lo en to cnjlo 

rH rH 1— i 



O 


lOvOO 


^r *s- o 


o o 


o 


"* cv) cni 


lo rsi to 


LO LO 





01 


in 


o 




en fH 


1— 1 


in o 




C O 


in o 


rH X! 




O in o 


C in o 


o o 




•rtHX 


OH£ 


O 172 




■POO 


•H O O 


X 




3 O CO 


•p o co 


O rH 




P X! 


3X 


in ea 




•H O rH 


P O i-t 


o 


in 


P CO cfl 


•iH t/5 CO 


rH -H 


c 


V) O 


P O 


C8 TJ 


o 


C rH -H 


l/IH'H 


o o 


H 


m n t3 


C W T5 


•H 2 


4-> 


o o 


MOO 


•o 


3 


O-H 2 


•H 2 


CD P 


P 


P T3 


OX) 


2 3 


H 


C8 o P 


•H O P 


o 


P 


> 2 3 


rH 2 3 


x! x 


Ul 


•H O 


X o 


p p 


C 


U&& 


3X£ 


■H -H 


M 


CL, P P 


0, P P 


s s 



rH i-HSS rH |S ^ rH i-l 

iH iH rH rH iH 

< < < << 



CO 

o 



CO 
4-> 
CO 



c 
o 



CO 

u 

3 

T3 

w 
o 



u 

(Li 
4-> 

c 

0) 



CO 

c 
o 



o 
o 
u 

3 

o 

CO 

ct) 



e 




r* 




o 




X 




CO 




l/I 




0) 




3 




c 




0) 




> 




o 




cs 




o 




ojr 




Qci 




T3 




CD 




(-. 




O 




l/I 




« 




O 




O. 




CO 




<+H 




o 




01 




c 




CO 




CD 




2 


/ — v 




o 


0) 


o 


X 


rH 


P 






II 


!- 




O 


r4 


Uh 


r^ 




cr> 


en 


H 


rH 




O >H 


X tJU 


e 




3 


U 


2 


O 




l+H 


X 




cu 


ul 


T3 


C 


e 


CO 


i — i 


a> 




S 


c 


v — * 


■H 






a> 


U) 




T3 


<D 


c 


rH 


O X 


rH 


CO 


H H 



rH 
O 

X 

6 
3 
Z 



ur) rH o r~- to to lo 

C7i ClOl!-~ o o o 



C 

o 
m -h 

O 4-> 

3 

0) H-> 
P,.H 
XP 

H en 

C 



cri <joio eti o r~- 
oi en cri oo o rH o 



o 
o 



oi 



o o o 
o o o 



o o o 
o o o 



LO LO LO 

cy> o\ o\ 



COO>sD 
CTi Gl en 



LO rH \0 

rH rH C-j 
O O O 



rH C-g OI 

O O cn 



tO rH O 
O O rH 



CTlCOH 

en a> o 



CTI CO 
Ol CO 



to o 

O CTs 



O O 
O O 



t-~ LO 

ai cti 



O vO 

o o 



o 


~£> vo O 


■* "* O 


o o 


o 


^rj- <v] csl 


LO CN1 tO 


LO LO 





in 


in 


o 




t/l rH 


rH 


in o 




c o 


in O 


rH X 




O en o 


C in O 


O O 




•iH rH X 


O rH X 


O CO 




+-> O o 


•H O O 


X 




3 O CO 


P O CO 


O rH 




+-" X 


3 X 


CO CO 




•H O rH 


+-> O rH 


U 


en 


■P CO CO 


•H CO CO 


rH -H 


d 


in o 


■P O 


CO T3 


O 


(S rH .H 


in rH -H 


O O 


H 


rH CO 'O 


C CO T3 


•H S 


fJ 


o a> 


rH O O 


*o 


3 


O -H s 


■H 2 


CD P 


4-> 


+-> T3 


o -d 


2 3 


H 


CO O +J 


■H CD +-> 


O 


+-> 


> 2 3 


rH 2 3 


X X 


tn 


■H O 


X o 


p p 


c 


rn XX 


3 XX 


•H .H 


H 


B, t- 1 +-" 


a. p +j 


s s 



S 3 



< < 



-27- 



Presence vs. absence of medical schools . Combining private and public 
institutions, universities with medical schools seemed to fare somewhat 
better than those without them relative to level of R&D funding throughout 
the entire period from FY 1969 to 1974. But such a combination produces 
misleading results, as the figures in Table 9A show. Private institutions 
with medical schools do show negative growth after FY 1972, but the extent 
of it is far less than for universities without medical schools (the drop 
in index numbers for the latter is from 100 in FY 1972 to 77.9 in 1974). 
In the case of public institutions, on the other hand, those with and those 
without medical schools both show substantial increases in R&D funds 
after FY 1972 — with growth over the two-year period to FY 1974 being 
roughly equivalent for the two groups. 

Trends in R&D revenues by Carnegie Commission categories . It will be re- 
called that total E&G revenues showed an average decline for private institutions 
after FY 1972, but that all of it occurred in the top category, Research Univer- 
sities I. Similar comparisons were made for R&D revenues; but the figures in 
Tables 10 and 10A on the following page do not agree with the E&G results. In 
general, all Carnegie Commission categories of private institutions had declines 
in R&D revenues after FY 1972, the index numbers for FY 1974 being as follows: 
Research Universities I, 90.9; Research Universities II, 94.1; Other Categories, 
90.3. 

In the case of public universities, there generally was substantial positive 
growth in sponsored R&D funds after FY 1972, with only Research Universities II 
showing a rather sharp decline from an index level of 107 for FY 1973 to 96.6 
for FY 1974. 

The comparative percentage changes in R&D revenues from FY 1969 to 1974 for 

the Carnegie Commission categories may be summarized as follows: 

Private Public 

Research Universities I 
Research Universities II 
Other Categories 



-12.3% 


6.7% 


-17.0 


-4.0 


- 3.7 


12.6 



-28- 



l/l 




0) 




•H 




P 




O 




DO 




01 


•» 


P 


P 


03 


o 


u 


p 




co 


a 


r-H 


O "P 


•H 


(U 


CO Q 


to 




•H 


Q 


S 


uff 


e os 


o 




U X 




1— 1 


« 2 


•H 


i — ' 


bO 




V 


l/l 


C -O 


^ 


c 


03 


CO 


U 


CO 




3 


X o 


XI X 




H 


l/l 




<u 


C 


3 


•H 


C 




cu 


to 


> 


P 


a 


CO 


06 


i— I 




rH 


Q 


O 


u!7 Q 


OS 






P 


X) 


c 


1) 


CD 


H 


P 


O 


10 


to 


c 


c 


o 


o u 


p. 


1 


m 


1 




toco 


c 


C<~* 


03 


o © 


CD 


•H O 


S 


P i-l 




3 


e 


P 11 


■H 


•H 




+-> "3" 


l/l 


to \o 


T3 


C Ol 


C 


I-H i-H 


V 




p m > 


H 


O (i, 



"5t 

Ol 



CO 

r- 

Oi 



p 

c0|oi 
V 



u 

CO 
•p 

til oi 



o 
r- 

Ol 



01 

o 



p 
co 

XI 

e 

3 
2 



o 

<4-t -H 

o P 

3 

CD <4-> 
0.-H 
>sP 

H CO 

C 



i— i r-^ cm o 

OlOiOt 
Ol Ol o^ o^ 



O LO O CM 
CM LO t— ( IO 



OOtN 


vOl/1 l/)N 


I— I CM 


i— 1 


^*o* 


h-o vor- 


vo lo r-- o 


\ONH3iO 


CMU1HH 


ulMvOM 



HH Ul Kl 
i-H CM 



O O LO CM 



*i- cm t^ lo oooin 

ClOCMN OOOCMrt 



HIOlflM 



•a- oi ■* r~- 



O ^ LO CM 



ONON 



i-H i-l 


lo 


cm 


lo *a- 


LO 


CM 


rH cm 






i-i 






W3- 












oi r~» 


tO 


oi 


oo o 


LTl 


Ol 


cm t~- 


vO 


r- 


oi CO 


o> 


o 


i-H LO 


LO 


co 


o r- 


*& 


CM 


cm to 


i_o 


cm 


o «a- 


LO 


CM 


i-H CM 






i-H 






■€«■ 












i-l "* 


LO 


cm 


vO CNI 


CM 


CM 


i-H O 


I«- 


lo 


•<*• co 


O 


i—i 


^- CO 


Ol 


O 


Ol LO 


tO 


rH 



CM to LO tO 
i-l CM 



vOOlNl/1 
**■ i—l i— I i—( 



LO ■* LO CM 



•<3- CM 00 «» 
LO i— I i-l CM 



to 




CO 


C co co 


CO 10 t/1 


+J 


O 4) CD 


C <D 4) 


c 


•H -H 'H 


O -H -iH 


CD 


P P p 


• H P P 


u 


3 -H -H 


P .|H -H 




p in i/i in 


3 10 10 CO 


i-H 


• H P t_ 4) 


P P p CD 


CD 


P CD CD -H 


•H CO CU -H 


C 


10 > > P 


•P > > Vi 


O 


C -i-l -H O 


CO -H -H O 


•M 


i-i (3 C QO 


CCC M 


■P 


3 3 1) 


i-h 3 3 CD 


CO 


(1) 4-i 


•»-> 


2 


p x x co 


U XX CD 




CD o u u 


•H O o u 




> P P 


i-l u u 




•h o3 c9 P 


X CD CD i-i 


CD 


1m 0> 0) CD 


3 0) CD CD 


U 


&, co co X 


a, co co x 


u 


D D P 


CD CD +J 


3 


rtttttO 


hq!b:o 


o 


!-H 


rH 


w 


< 


< 


03 



u 



ctf 
P 
C7D 



c 
o 



CD 
U 
3 
T3 
W 

P 
O 

4-1 



o 

X 

to 

co 

CD 
3 
fl 
CD 
> 
CD 
Oi 

Q 

a: 

CD 

p 
o 

CO 

c 
o 
p. 

CO 

LM 

o 

co 

c 

cO 

CO ^— > 

2 o 

o 

CD i-H 
X 
p II 

p CM 

O t- 

>p a> 
i—i 

(0 

p >- 

CD U-. 
X 

S p 

3 O 
2 <P 

X co 

CD C 

73 CD 

C CO 

i-i S 

c 

•H O 
i-H 
CO 

T3 CD 
C i-H 
CD X 
P CO 



u. 



CM 






o 

CTi 



P 
CD 
X 

E 
3 



C 

o 

CP -H 
O P 
3 
CD P 
CX-H 

>sP 

E-h io 
C 



NOIHM 



CO i-H vD O 



r-( O "3- O 
C7) tTi d Ol 



CO O O ^3- 



ho LO r~- i— i 

OOOIH 



■q- CM O CO 



<r toco io aiot^i— i 

OICTIOIOI O t— I O rH 



O O O O 



o o o o 



o o o o 
o o o o 



co LO rf to 



o o o o 
o o o o 



•3- O CM OO 



LO LO i-H rH 

Ch en <o> o\ 



LO <* Tt ^t 



aih-oa 

Ol CI O Ol 



(Ococooi 



i-H CM LO CO 
O O O OO 



Ol MON 



rH Ol tO tO 

O Ol o o 



to to to to 

O O i-H Ol 



Ol CO O Ol 
Ol Ol O Ol 



OOiCMLO <=i-cmco^i- 

*3-i— IrHi-H LOi-HrHCM 



CO 




C co co 


CO CO to 


O CD CD 


C CD CD 


•p -H -H 


O -H -H 


p p p 


• H *J +J 


3 -H -H 


p -H -H 


P CO CO CO 


3 io io co 


■H p p CD 


P p P CO 


P CD CD -H 


•H CD CD -H 


CO > > P 


P > > P 


C -H -H O 


CO 'H 'H O 


i-H C C OO 


C C C DO 


3 3 CD 


hddhj 


CD P 


p 


P X X CO 


U X X co 


I4UOU 


■H L> U C_> 


> P P 


rH p p 


•H CO CO P 


X CD c8 P 


P CD CD CD 


3 CD CD CD 


Oh CO CO X 


CL, CO CO X 


CD CD P 


CD CD P 


i-H OS o-: o 


i-H OS OS O 



-29- 



Federally Sponsored R&D Revenues 

Trends in federally sponsored R&D revenues, by type of institution, are 
shown on the following page in Table 11 (means in constant dollars) and Table 
11A (index numbers with the mean for FY 1972 as the base). The results will be 
discussed mainly in terms of index-number trends shown in Table 11A, since such 
values are most easily compared within and among the several categories of 
institutions. 

There are significant differences between the trends for federally spon- 
sored R&D revenues and those for total R&D revenues shown earlier in Tables 9- 
9A. For example, federal R&D funds stood at a higher index level in FY 1969 
for all institutions combined than was the case for all R&D revenues (106.2 
vs. 102.1). But federal R&D funding level declined more sharply overall from 
FY 1969 through 1974 than happened for all R&D revenues - 5 (-9 .7% vs. -6.1%), 
although the terminal index differences were not great (94.5 vs. 95.9). 

Private vs. public universities . The private institutions showed declines 
in federal R&D funding levels after FY 1972 somewhat greater than those 
found for all sponsored R&D revenues, but the patterns and the end results 
in FY 1974 were essentially similar (90.7 vs. 91.2 in index values). 

Public institutions showed growth in federal R&D funding after FY 1972, 
but the FY 1974 levels were lower than for all sponsored R&D revenues (101.2 
vs. 103.8). Both types of institutions had appreciable declines in federal 
R&D funding from FY 1973 to 1974, and the public sector showed the greater 
rate of decline. But for FY 1974, the index number for the public group 
was 101.2 while that for the private group was 90.7 — relative to the base 
year (FY 1972 = 100). 

The overall percentage changes in federal R&D funds from FY 1969 to 1974 
were: private universities, -17.1%; public universities, 0.5%. 

Presence or absence of medical schools . There appear to be no signifi- 
cant differences between the trend patterns for federally sponsored R&D 
revenues and those for all sponsored revenues, in terms of presence or 
absence of medical schools — when public and private data were combined 
in each case: institutions with medical schools showed less decline in 
both total and federal R&D funds than those without medical schools. 



-30- 






H 


r~- 


03 


Oi 


CD 


iH 


>- 




iH 




cd 




u 




U} 


iH 


•H 


t-- 


U. 


sn 



r< 



c 
o 

■4-1 -H 
O 4-> 

3 

0) 4-> 

CVH 

X4-> 

f-i oi 

C 



i-H 


Nl/IM 


o 


o «* 


\o lo 


cm 


OOtTt 


oo 


o o 


"* cn 


r^ 


oo oo o 


oo 


en "» 


oi ■* 


o 


ooHm 

I-H 


•»3- 


t^ CM 


Ol to 


w» 










CTl 


r-~ r-~ r-t 


o 


O0 >—l 


O !"- 


r-l 


r-~ CTl LO 


oo 


<xl o 


** CTl 


O 


i-l r-l fNl 


i-H 


to \o 


tO tO 


t-. 


OlNl/1 


LO 


CO CM 


O to 



rH CTl "O - I~- 

i-H SOOO 



CM 


o 


cn 


■* LO 


CM 


r- 


LO 


Ol cm 


CO 


r- 


*t 


Oi cm 



as 


















r^ 


CTl 


i— 1 


vO 


"3- 


r-. 


cm 


CTl 


•a- 


■&<3- 


















to 


o o o 


Oi O Ol 


00 


CO 


LO 


OOl* 


r- 


\0 •■o 


LO 


<* 


oo 


«3- 


LO 


lo 


o 


lo 


tO 


o 


O 


o 


oi 


l— I 

rH 


o 


•*i- 


t— 


CM 


Ol 


«* 


<»>> 


















lo 


LO 


cm 


i/i 


i-H 


r- 


*3- 


to 


*o 


o 


oo 


to 


to 


LO 


*t 


LO 


0O 


cm 


«* 


CM 


vO 


cm 


Oi 


a-i 


LO 


to 


«* 


r-- 


o 

1—1 


CM 

i-H 


t>» 


«* 


I-. 


CM 


o 

i-H 


^T 


■5r> 


















CTl 


OOl^ 


vD 


LO 


r- 


"* 


■3" 


"* 


rH 


Ol 


to 


LO 


LO 


r— 1 


fO 


UD 


LO 


t~- 


oi r- 


00 


o o 


"3- 


^O 


n 



















«r> 



o 


o o o 


«* ■* o 


o © 


o 


■r)- CM CM 


LO CM tO 


LO LO 





01 


oi 


o 




01 i-H 


rH 


in o 




c o 


oi O 


rH ,C 




O in O 


G oi O 


o u 




•H i-H J- 


O i-H X 


o to 




4-> O U 


•H O U 


X! 




3 O CO 


+-> o to 


O rH 




+-> X 


3 XI 


to a) 




•H CJ rH 


*J UH 


o 


01 


+-> CO CO 


•h to nl 


rH -H 


e 


ol O 


+-> o 


n) '"O 


o 


C >H -H 


UlH-ri 


O <U 


H 


H CO'O 


C « T) 


•h 2 


■u 


U 4> 


H DID 


-d 


3 


<U -H 2 


•rtS 


0) 4-> 


4-> 


+J n3 


U T3 


2 3 


H 


n) 0) *j 


•ri t)*J 


o 


*J 


> 2 3 


H23 


X J3 


to 


■H O 


^2 O 


4-> 4-1 


C 


»- xx 


3 .CC 


•H -H 


i — i 


CL, 4-> 4-> 


O. +J +J 


? s 



< < 



u 



a) 
4-> 

to 



c 
o 



CO 

u 

3 
T3 
PJ 

r. 
O 

Mh 

Sh 

a> 

4-> 

c 
u 



CO 

a 

o 



u 

M 

3 
O 

to 

CO 



cn 

<u 

3 
C 
CD 

> 
O 

eC 

Q 

Oi 

a> 

!h 

O 
to 

C 

o 
p. 
to 



CO 




u 




01 




X) 




o 


^— s 


UL, 


o 




o 


l*H 


r— I 


o 






II 


1/1 




c 


CM 


CO 


P- 


a> cr> 


2 


^H 


CO >H 


X u. 


+J 






Ih 


1-4 


o 


O tw 


m 






U1 


in 


C 


h 


CO 


0) 


cu 


Xi 2 


S 


- — ' 


3 




2 


i-H 




i—l 


X 




CD 


<L> 


T3 


^H 


C XI 


l — l 


CO 




H 


C 




•rH 


C 




■H 


1/1 




T3 


c 


C 


s 


4> 


o 


(h JC 


E-i to 





CXI 


r- 


r-- 


CO 


CTl 


<u 


rH 


>H 




rH 




CO 




U 




CT; 


rH 



u 

o 
X 

e 

3 
2 



c 

O 

MH -H 
O 4-> 

3 
0) +J 

CVH 

c 



CTl 



o oo to 



rH rH O 

o o o 



a> csi 
cn oo 



CO 


tO rH VO 


r~- r— oo 


to r-- 


CTv 


OlON 


o o o 


O CO 



o 

o 



o o o 
o o o 



o o O 



o o o 
o o o 



o to to 



to cn lo 



\0 CTl C-~ Cxi OOOvD 

O O O rH CD CTi O 



rH HSS 



s s 



o o 
o o 



vO 


\Q \0 LO 


i^ r- o 


vO lO 


CTi 


CTi CTi CT^ 


O^ CT^ CT> 


o^ a^ 



CM 


O CTl CM 


NNOl 


Ol CO 


■* 


lO >y lO 


CM CM tO 


to ^f 


o 


o o o 


o o o 


o o 



** o 

O rH 



o 


o o o 


«* -* O 


o o 


o 


«d" CM CM 


LO CM tO 







1/1 


01 


O 




C/l rH 


rH 


ol O 




c o 


Ol O 


rH X 




o to o 


C 01 O 


O U 




•H ^-! jC 


OH£ 


o to 




4-> O O 


•H O U 


X 




3 O to 


4-> O tO 


O T-t 




4-« x; 


3 X 


to n) 




■H U rH 


*J UH 


L) 


</> 


4J to nj 


•h to n) 


l-H -H 


c 


to O 


4J U 


n) -3 


O 


CHH 


Ol rH 'H 


O CD 


H 


ih CO -3 


a CST3 


•H 2 


4-1 


u a> 


H U 11 


■3 


3 


a> -h 2 


•h 2 


CD 4-» 


4J 


4J XI 


O T3 


2 3 


H 


nj <u 4-> 


•H CD 4-> 


o 


+J 


> 2 3 


rH 2 3 


X X 


10 


•H o 


X O 


4-1 4-1 


G 


h x; x: 


3 X X 


•H -H 


HI 


Pa +J 4-1 


Ph 4J 4J 


2 £ 



< < 



-31- 



Federally sponsored R&D revenues by Carnegie Commission categories . 
Federal R&D trend data for institutions classified by the Carnegie Commission 
categories are shown in Tables 12 and 12A on the following page. The compara- 
tive percentage changes over the six-year period (1969-1974) for public and 

private institutions were as follows: 

Private Public 

Research Universities I -17.5% 1.0% 
Research Universities II -20.2 -6,0 
Other Categories - 7.2 5.8 

The patterns of changes shown in these figures for federal R&D funds are 
quite similar to those described earlier for all R&D revenues. But the magni- 
tudes of the percentage decreases are greater for federal than for total R&D 
funding in private universities; and in the case of the public universities 
showing increases, they are less for federal than for total R&D revenues. 

In general, the private universities have suffered serious losses of 
research support in recent years — nonfederal as well as federal — with Research 
Universities II being hardest hit but with Research Universities I also showing 
a sharp decline. And it was the latter, it should be recalled, that showed a 
substantial drop in total E&G revenues after FY 1972. 



-32- 



cO 



o 
o 



c II 
o 

•H «* 
tO sO 
to CTl 

•iH rH 

6 

O P-. 

u 
CD r< 

■H O 
Q0 4-I 

0) co 

CS r-i 

rH <+H 

co o 
u Q 

>,Q 

X UJ 
PS 

to 

<D X 
3 hh 
C Z 

0) s— ■ 

> 

<d 



o 

OS 

-O 

0) 

o 

en 
C 
O 

a 
to 



>srH 



c 

03 



a) 

(!) 
T» 

0) 4-> 

Uh in 

c 

c o 

co u 
a> ■ 
2 ' 

to 

C CD 

•H -H 

in O 

xi oo 

c <u 

*-> 

U CO 

H U 



CD 

rH 

X 

CO 

E- 



CTl 



u 


r- 


cd 


Ol 


cu 


— I 


>~ 




rH 




rt 




o 




10 


t— H 


•H 


r-- 


P-, 


en 



a. 

sO 
CTi 



cu 

XI 
S 
P 

Z 



c 

o 

m -h 

o 4-> 

3 

CD ■(-> 

d,-H 
>.4-> 

f-H tO 
C 



csi r-^ *3- 


^r 


o 


o 


t~~ «* 


COlOH 


ct> 


co 


o 


'd- CTi 


CO <— 1 O 


CM 


00 


to 


o r-s. 


CO T-~ ^t 


rsj 


^r 


CM 


^ rH 


i— 1 






rH 




■fog- 










r- oo \0 


to 


o 


sO 


00 sO 


r~~ lo <* 


to 


oo 


o 


sO Cs] 


rH t- i-H 


to 


•-< 


rH 


to 00 


At M A 


•* 


M 


r- 


•> * 


oii^q- 


Csl 


LO to «* I— 1 


I—I 






rH 




<r> 










i/icot 


rH 


Cs] 


CTl 


CTi to 


CTi r~- r^ 


O 


cni 


r-j 


sO 00 


t~- CTi to 


lo 


oo 


cr> 


Csl .£> 


cm oo •* 


Csl 


"* 


rH 


<sf r-H 


rH 






r— 1 




H=r> 










lOOON 


Csl 


Oi 


C-J 


r-» oo 


OOllfl 


rH 


r-. 


to 


LO kO 


•*HIO 


to 


sO 


to 


C-J \0 


CTi oo "* csi 


•"* 


^ 


"* rH 


rH 






i-H 




09- 










lo lo csi 


fNl 


r-> 


rH 


C%1 \D 


oo 00 -* 


LO 


LO 


O 


rH LO 


csj cti oo 


to 


<y> oo 


LD CO 


OOl"* 


csi 


■=r 


rH 


«* rH 


l-H rH 






rH 




*r> 










O O O 


tO 


vO 


ro 


r-\ O0 


rH O tO 


r^- 


LO so 


c-1 rsi 


NOOO<* 


00 


O 


Csl r— 


O O LO 


CJ 


*rj- Csl 


"st t-H 


rH Csl 






r^ 




«r> 











O CTi Csl LO 

tHHH 



10 

CS co to 
o a) cd 

•H -H -H 
4-) 4-> 4J 

3 -iH -H 



10 l/l 
U CD 

cu ■ 



rH C 

CD 

+-> XI 

CO (J 

> !-» 

•H a) 

H CD 

O, to 
CD 

rH OS 



O 

00 



■sf Csl O0 
LO f-H rH 



Csl 



CO CO CO 

C CD CD 
O -H -H 

■H4J(J 



3 



■M > 



CO 1/1 CO 

S-c f-i CD 

CD CD -rt 

> U 

H O 

C oo 



X! cO 
U U 

CO in 
CD CD 

CD +-> 
OS O 



H3D (D 

+J 

O X X CO 

•H U uu 

rH J-c S-i 
X CO CO r. 
3 CD CD <D 

a. co to x 

CD CD 4-> 
rH OS OS O 



to 
U 



c 
o 

■H 
4-1 

03 
O 

3 

W 

I-H 
O 

LH 
I- 

(D 

■M 
C 
CD 



03 
CS 
O 



C3 
2 



CD 
CJ 

r, 

3 
O 

to 

cfl 



to 
CD 
3 
£ 
CD 
> 
<D 
OS 

a 
us- 
es 

CD 

r. 

o 

to 

C 

o 

o. 
co 



rt 




Vh 




n> 




CD 


, , 


Uh 


o 




o 


14H 


<-< 


o 






II 


tfl 




c 


C-] 


a 


r~ 


CD 


CT, 


2 


r-H 


CD >- 


XPL, 


4-» 






Lh 


)h 


o 


O Mh 


C+H 






(At 


l/l 


C 


I- 


rt 


CD 


o 


c s - 


3 




Z 


Csl 




rH 


X 




CD 


CD 


T3 rH 


C X 


t—i 


rt 




H 


c 




■iH 


c 




•H 


to 




T3 


c 


C 


s 


CD 


o 


rH X 


H CO 





CM 


r- 


L - 


rt 


CTi 


CD 


rH 


^ 




r-H 




rt 




u 




to 


rH 


■H 


r^ 


u. 


CTi 



o 

CT 



Lh 

CD 

X 

S 
3 

z 



c 
o 

LH -H 

o +-> 

3 

CD •!-) 

>s+-> 

H "l 

C 



r~- «rjs oo r-s- 



o o <-\ t-I 
CTl CTi CTl CTi 



r-~ so co to 



Csl rH CO sO 



rH tO .3- O 
CD O CTi O 



^T CTi to LO 



to to m- to 

CTi CTl CTl CTl 



OOOO 



tss CTl Csl 00 
OOOO 



OOOO 



OOOO 
O O IO O 



O CTi LO ^f 



O O O O 
OOOO 



O O !--. rH 



sO LO CTi Csl 
CTl CTi CTl CTi 



O to fs- rH 



l-^ LO CTi CTl 
CTl CTi CTi CTi 



t-» CTi r— to 



l/llOOst 
O O rH CTi 



tO sO O CTl 



Csl 00 LO O 

O CTl O 1— I 



I-sH oiIn 



CTl CTi LO CO 

O O rH CTl 



O rH OO Cs> 

O O CTl O 



vOCTiCslLO "rj-CsloO-* 

Tl-rHrHrH LOi-HrHCs] 



to 






C to t/> 




to to to 


O CD CD 




C CD CD 


•H -H -H 




O -H T-I 


4-> 4-1 +-> 




•H +-> 4-1 


3 -H -H 




4-) -H -H 


P to to 


to 


3 to 10 co 


•H !h V-i 


CD 


4H rH rH H) 


+J CD CD 


■r-t 


•H CD CD T-I 


m > > 


u 


4-> > > U 


C -H -H 





tO tH T-I O 


w C CS 


Ofl 


C CS c 00 


3 3 


CD 


h3D 0) 


0) 


+J 


4-> 


■M rC X 


CO 


UX X CO 


CO U CJ CJ 


■H u u u 


> f-i tn 




^ (-H rH 


•H CO CO 


rH 


X CO CO rH 


rH CD CD 


(D 


3 CD CD CD 


Oh <o to X 


Oh to to X 


CD CD 


+J 


CD CD 4-1 


rH OS OS O 


rH as os 



-33- 



V. TRENDS IN EXPENDITURES FOR BIOMEDICAL-BEHAVIORAL RESEARCH 

All of the financial data reported in the preceding sections were revenues 
from the NCES-HEGIS financial survey, which provided only aggregate data for 
the entire institution and all academic fields. In order to study R&D expen- 
ditures in the biomedical-behavioral fields, it was necessary to secure data 
from the National Science Foundation's "Survey of Scientific Activities of 
Institutions of Higher Education," which collects data on current expenditures 
for separately budgeted R&D in the sciences and engineering. This survey has 
been conducted biennially from FY 1964 through 1972 and annually thereafter. 
However, no records for FY 1966 were available, with the result that the present 
report covers only the fiscal years 1964, 1968, 1970, 1972, 1973, and 1974. 



Prior to FY 1973, this NSF expenditure survey asked institutions to exclude 
"development" funds from their reports and was described as a survey of 
"research" expenditures. But beginning with FY 1973, composite reports of all 
"research and development" funds were requested. In addition, institutions 
were asked to make overall percentage estimates of expenditures for basic 
research, applied research, and development. For FY 1974, NSF reported that 
about 4 per cent of all expenditures for academic R&D was estimated to be 
for "development." 

The 1973 change in the scope of the NSF survey raises a serious question as 
to the comparability of prior data with those collected after the change. With- 
out any evidence as to whether or not the institutions were actually excluding 
"development" expenditures from their earlier reports, it has been decided to 
use data for 1973 and 1974 as reported rather than to reduce them by, say, the 
estimated average percentage of development funds. It is probable, in any 
event, that most of the latter type of funding occurred in the fields of 
engineering and physical sciences. Although no data are available to support 
the assumption, it seems unlikely that universities would have received an 
appreciable proportion of their federal funding in the biomedical-behavioral 
sciences as "development" grants. Furthermore, "development" activities 
would probably have been included in grants made primarily for research pur- 
poses and hence reported for "research" in the survey. 



-34- 



The term "biomedical-behavioral" refers here to a composite of the follow- 
ing four groups of disciplines included in the NSF "Survey of Scientific 
Activities of Institutions of Higher Education": (a) biological sciences (in- 
cluding agriculture), (b) clinical medical sciences, (c) life sciences not else- 
where classified, and (d) psychology (all fields). 

A total of 143 institutions of the 148 in the original sample supplied 
data for the NSF expenditure survey, for all six years indicated above. By 
contrast, it will be recalled that for the NCES financial surveys the number 
of institutions providing data every year was only 100. 

Since the four components of the "biomedical-behavioral" complex were 
found to differ considerably among themselves in magnitude of mean expenditures 
and in trends, it seemed desirable to present results both for the composite 
category and for the individual discipline groups. The analysis and presenta- 
tion in the case of the composite category followed the general plan used in 
the two preceding chapters — for both total and federally funded R&D expendi- 
tures: (a) pairs of tables showing means and index-number trends for all 
institutions combined, for all private institutions (with and without medical 
schools) , for all public institutions (with and without medical schools) , and 
for all institutions with and for those without medical schools; (b) parallel 
pairs of tables showing means and index-number trends for private and public 
institutions classified by categories of the Carnegie Commission on Higher 
Education. 

In the case of the four discipline groups, it was decided to present data 
in these two analytical formats only for federally funded R&D expenditures. 



-35- 



Th e reasons were partly to avoid undue proliferation of tables and partly 
because federal funds constituted both the greater part of R&D funding in 
these fields and the main focus of interest in this study. 

In order to simplify comparisons of data within each of the two modes or 
formats for classifying institutions, all of the tables based on Carnegie 
Commission categories have been placed in Appendix B. The other set of tables — 
i.e. those with breakdowns by type of control and medical-school status — will 
appear in the text (each one following the page on which it is first cited) . 
R&D Expenditures for All Biomedical-Behavioral Sciences 

All sources of R&D funds . Mean expenditures for all biomedical-behavioral 
expenditures by type of institution, in constant dollars, are presented in 
Tables 13 and 13A on the following page. It is evident from inspection of 
the index-number trend shown in the first line of Table 13A that total expen- 
ditures for biomedical-behavioral R&D have followed quite a different pattern 
from that shown for all sponsored R&D revenues in Table 9A, Biomedical-behav- 
ioral funding increased substantially after FY 1972 (the base year) , whereas 
all sponsored R&D revenues declined. 

Private vs. public institutions . Although the positive growth rate is 
somewhat lower for private than for public institutions, the trend is 
definitely positive from FY 1964 through FY 1973 — with a slight "recession" 
in FY 1974. 

The pattern for public institutions differs slightly from that for the 
private sector, in that a very considerable increase was recorded for 
FY 1970 which didn't occur for private universities. Furthermore, the 
growth rate after FY 1972 was considerably higher for the public than for 
the private universities. But all were substantially positive and both 
groups showed moderate declines in FY 1974 from the peak year, FY 1973. 



-36- 



x 
u 
u 
a 

0> Q 
in us 

o> at 
d 

X 

rH rH 

cd Z 

U w 

o 

•H 1/1 
>T3 

03 C 
X CO 

pa 3 
i o 

co H 
o 

•h e 

13 -H 
0> 

e m 

o rH 

•h co 

« i-H 

rH 

<-< o 

rH Q 

< 



n) 



O 

4-i 

in 
m C 
co O 
rH Urt 
3 ' /-> 
4-> l o 
•H C O 
T) OH 
C -H 
CD 4-> II 
0.3 
X 4-> "3- 
H-HO 

4-> Oi 
C WH 
cO C 

C O « 

•H rH 

<d o 
in cu+J 
T> >> cfl 

C E-> iH 
CD 4-1 
rn >s <D 
H X Q 






"3- 






co|cn 

>- 



o 
cn 



00 






rH 

s 

3 

z 



c 
o 

4-1 -h 
O +-> 
3 
0> 4_> 
D.-H 

>s4-> 

H w 
C 



to 

r- 

■bO- 

CM 
CM 

co 

LO 

■to- 

CM 

CM 

in 

"3- 
00 
CM 

lO 

O 

"3- 
o 

LO 



00 

cn 



to 

*r> 



to 



OOIO 

cn r-- o 

tor-.^- 



LO \0 l/l 
tJ- cn CM 
t-1 0> "1 



tO 0> CM 
lO 0> 00 
O CM CM 



W 

c 
o 

•H 
3 



in 

C 



O Cn i-l 



to oo oo 
cn cn r- 

ON H 



"3- Tf tO 

O r~- cm 



to o i-l 

VO f rH 
CM VO 00 



^- CM CM 
U"> tO CM 



in 

in t-h 

c o 

O in o 

•H rHX 
■4-> O U 

3 O W 
4->X 

•H O rH 

+j co ce 
m o 

c 



CO , 



wo 

u oi 

S 

4-> TJ 

Cfl CD 4-> 

>2 3 

•H O 

0. 4-> 4-> 



O O CM lOOl 

*3- o r~- o \o 

tO CM lO <y> Cn 



\C Cn rH W> 00 to 



tO CM *3- 
00 LO CM 

ki m<* 



LO oo 
CM 00 



toooi 

OlH<f 
f- o o 



tO "3- 
to ^- 
•3- LO 



vo r- r~- 

tO \0 "3- 



Ol "3- LT> 

00 to LO 



in 

rH 

in o 
C m O 

O <-> x 
•h o o 
+J Otn 
3X1 

4-> O rH 

•H t0 CCj 

4-> O 

in rH -H 

C Cfl 13 

►h U 0> 

■h 2 

•H <1> +J 
rH 2 3 

X O 

O. +-> +-> 



^O CM 

cm r-^ 



to to to r- vo 

"3- t-- f- rH rH 



OlKII-- OO CM 

O CM CJi LOtv 

oo rH cri oo to 






m 

i— i 

o 

m o 

rH X 

o o 

O C/) 

x: 

O rH 

co rt 
o 

rH 'H 
CU 13 

O CO 
•H2 

•o 

ID 4-> 

2 3 
O 

rC X 



s > 



rH HJ3 H%% rHrH 

rH rH rH rH rH 

< < < << 



c 
o 



ca 

T3 

c 

3 
O 

u, 

It) 
o 
n 
<u 

■H 

u 

CO 



CO 

c 
o 



CO 

u 

■H 

13 
<L> 

e 

o 



o 
o 



rH 

O 



in 

« CM 

3 o> 



ID 
X O 

W IH 
rH 10 

< CO 
co 

O ■_/ 

in to 

C rH 

CO 
co 

2 rH 
rQ 

CO CO 

rC H 

c 

M -H 
O 

M-i C 

m O 

rH rC 

CO CO 

ex 

3 O 

Z r- 

CO 

X CO 

co in 
13 co 

C Oi 



cd 

O 



in 
i3 

C 



CO X 
(h CO 



CO 


< 


z 


to 




rH 


. , 


CO 


CO 


rH 


u 


J3 


rH 


CO 


3 


H 


O 




</) 




CO 





CO 

o 

CI 



rH 

0) 

6 
3 
Z 



c 
o 

■+H -H 
O 4-> 

3 
CO +-> 
P.-H 

X4J 

E— in 

C 



cn mintn ■— i r-. i~» 

O O O rH rH O rH 



to to to 

CM CM CM 



o oo t^ cn 

rH O O rH 



O 
O 



o o o 
o o o 



o o o 
o o o 



O "3" ^l- rH LO cm cn 

o cn cn cn o o o 



LO 

cn 



lO 



•* to cn 
cn cn cn 



cn 0O rH 



O rH tO 

r^ r-~ >o 



cn t~^ co 

t-\D Ol 



O rH 



cn to 

O rH 



O O 
O O 



co r» 
cn o 



cn >* 

00 rH 



cn to 
*o cn 



to 


»3- CM CM 


cn *3- io 


o t~- 


>* 


U") tO CM 


oo to to 


o r^ 





m 


in 


O 




in i— i 


rH 


in O 




C O 


in o 


rH X 




O in o 


(3 in o 


o o 




•H rH J2 


O rH X 


O CO 




4-> O O 


•H o U 


X 




3 O CO 


■M O CO 


O rH 




<-> -C 


3 rC 


CO CO 




■H O rH 


4-1 U rH 


u 


in 


4-> CO CO 


•H CO CO 


rH -H 


C 


in u 


4-> O 


CO 13 


O 


fi rH -H 


in rH -H 


O CO 


rH 


H Ufl 


C CO 13 


•H 2 


t-> 


O CO 


HOD 


13 


n 


ID -H 2 


•h 2 


CO 4-> 


4-> 


+-> 13 


U 13 


2 3 


H 


CO CO 4-> 


•H CO 4-> 


o 


»J 


> 2 3 


rH 2 3 


X X 


in 


•H O 


-Q O 


4-> 4-> 


(3 


rn X X 


3 XI XI 


•H -H 


— t 


CX, +-> -M 


CL, 4-> 4-> 


IS S 



< < 



-37- 



Presence vs. absence of medical schools . The index-number trends presen- 
ted in Table 13A show that institutions without medical schools — both 
public and private — had markedly higher rates of expenditure increase for 
biomedical-behavioral R&D than those with medical schools. For example, 
private institutions with medical schools had an index of 105.2 for FY 
1974 in comparisons with 113.9 for institutions without medical schools. 
The corresponding figures for public institutions were 107.6 and 117.2. 

Combining both private and public institutions, those with medical schools 
had an FY 1974 index of 106.3, while for those without them the index was 
116.7. It should be noted, however, that for both the public and private 
sectors, institutions with medical schools had far higher expenditure 
levels in real dollars than those without medical schools — as shown in 
Table 13. For example, in FY 1974 private institutions with medical 
schools had mean expenditures of approximately $9.8 million, while for 
those without medical schools the corresponding figure was $1.5 million. 
The disparity was less for public institutions: mean expenditures of 
$8.2 million for institutions with medical schools in FY 1974, in contrast 
to $3.6 million for universities without medical schools. 

Differences among Carnegie Commission categories of institutions . The 
trend data for all sources of R&D funds by Carnegie Commission categories 
are shown in Appendix TABLES B-l and B-1A. It is important to recognize 
that Research Universities I, both public and private, have by far the 
largest share of the R&D funds — as shown by the means in TABLE B-l. 

Even so, private Research Universities I had considerably higher per- 
centage gains between FY 1964 and 1974 (62 per cent) than either of 
the other two categories — which in descending order had gains of 11 
per cent and -11 per cent during that period. 

For the public universities, the 11-year increases were: Research Uni- 
versities I, 42 per cent; Research Universities II, 32 per cent; Other 
Categories, 50 per cent. 

Federally funded R&D expenditures for biomedical-behavioral sciences . 
Data comparable to those in Tables 13-13A for all biomedical-behavioral R&D 
expenditures are presented in Tables 14 and 14A on the following page, for 
federally funded R&D expenditures in these fields. 

The trend in federal R&D funding for all institutions combined differs 
somewhat in pattern from that for all biomedical-behavioral R&D funds; but the 



-38- 



Q 

i-H CIS- 
CO OS 
r. 

O X 

•H 1— I 

> 2 

co w 
X 
cd in 
m-o 
■ c 
i-i co 
nj in 
U 3 
•H O 
13X 
0> H 
E 
o c 

•H -H 

oa 

in 
U U 
O co 



en O 
0) Q 
(-. 

3 4-> 
4-> C 
•h co 

13 4-> 

C in 
0) C 
CUO 

X u 

W ' 
I 

<u o 

TJ-H 
e +-> 
3 3 

Uh 4-> 
■HCO 

>s4-> /— N 
H 10O 
rH C O 

CO >-H i-l 

S-c 

CD <+H II 
13 O 

0) 1- 

IJU <D \D 

O.Ol 

C >N.-H 

co H 
cd >- 

2 5K Uh 

X 
C 

u 

« 

CD 



Ol 



tO 



h 


r~. 


03 


O. 


<U 


rH 


>•" 




r—l 




cO 




U 




in 


O 


•H 


t~~ 


Ph 


Ol 



r. 

e 

3 



13 

C 



h 

o 
4-J 

CO 



4) in 4H 

u o cd 
H OS Q 



XI 
ca 

H 



C 

o 

m -h 

O 4-> 

3 

4) +j 

Ph-H 
>n4-> 

E-i in 

C 



(N. 


COHH 


•=■)- CO CNI 


CNI Tj- 


to 


In. Ol tO 


i 10 m 


Ol t-^ 


t-N 


00 LO Ol 


O «* LO 


•^f to 


*• 


•* »s 


«k « m 


n ». 


to 


•* in 


to LO i-H 


O i-H 


■fcO- 








CO 


LO rH CO 


o to O 


CNI rH 


o 


«-)- H* t< 


i-H o r- 


CNI O 


CTl 


O CO o> 


NCOLO 


oo ■>-)• 


•a 


•\ *. 


M •* «\ 


«t OK 


tO 


lo in 


IO Ul i-H 


vO rH 


■t/> 








(VI 


cni to to 


lo h^- lo 


rH VO 


f- 


CNI O i-H 


tO rH vO 


00 to 


>* 


WOOl 


oo cni to 


O CNI 


•* 


» •* 


•X «k •» 


•S e. 


to 


"* IN 


NlflH 


VO r-i 


<*> 








VD 


o o o 


00 t^ o 


i-H "* 


\0 


CV1 Tj- UO 


"* "* o 


t~- to 


vo 


mow 


i-H LO O 


CNI •* 


•. 


«\ •. 


MA* 


■s #s 


to 


"* in 


K1WH 


VO rH 


«*» 








00 


i— I r-~ to 


r~ r^ •* 


O O 


oo 


i-l vO Ol 


o ^- Ol 


%o to 


LO 


NNOl 


oi woo 


oo *o 


«* 


•t *t 


■v •» •» 


M n 


tO 


>* in 


Cvl "* i-H 


LO rH 


«* 








Ol 


LO i-H CO 


LO LO tO 


CNI CO 


vO 


VO o o 


COOOl 


O vO 


vO 


<tnl~- 


i-H tO "* 


tO CNI 


•* 


* •» 


«' *\ M 


M M 


CNI 


tO LO 


<N1 tO i— 1 


** rH 


■&«> 









to 


«* CN) CNI 


a. ■* i-o 


o IN. 


»* 


LO tO CNI 


oo to lo 


\o in- 



C 

o 



3 



in 
C 



en 

1/) rH 

C O 

O in O 

•H rH XI 
■!-> O U 
3 O C/3 

■P X! 

•H O rH 
+-> C/2 a) 

in u 

H B)T3 

O CD 

4> -H S 

It DP 
> 2 3 

•H O 

r. X! X 

Ph 4-> -M 
•H -H 

^H S S 



rH 
C/) O 

C en O 

O rH X 
•H O U 
■M O CO 
3X, 

+J O rH 

•h w rt 
*-> u 
tn i-H -h 

C CO TJ 

f-h o a> 

O -3 

•HOP 
rH S 3 

X o 

3 X. X. 

a. +-> p 



tn 

o 
en o 

i-H XI 

o u 

O CO 

x: 

U rH 

CO CU 
U 

rH -H 
CO T3 
U CD 

•H S 

T3 

<D 4-1 

2 3 

O 

X X 



£ s 



S SB rH rH 

rH rH 
<< 



o 



CO 

c 

3 
O 

Uh 

CD 

u 
c 

CD 
•r4 

u 

CO 



CO 

c 
o 



CO 



CD 
O 
U 

3 
O 
CO 
CO 



o 
o 



in r^ 
1> en 

r. ,-( 

3 
*>>- 

•H LIh 

13 

C r. 

0) o 

PhCH 

X 
W in 

13 cO 
CD <d 

12 s 

C ^^ 
3 
m -a- 

rH 

>N 

•-H CD 

CO X 
U CO 
(D H 

<D C 

Uh iH 



4H C 
O S 

o 

in X 
C CO 
cO 

<D X 
2 U 

U 
CD CO 
X <D 
+-> in 

CD 

r. OS 

o 



CO 
in f-H 

r. O 

CD -H 
X > 

e co 

3 X 
Z CD 

pa 

X i 

CD rH 

•O CO 

c o 

H -H 

13 

c 0> 

■H 6 

o 

in th 
13 pa 
C 

CD ri 
r. O 
H C|H 



o 



r-COH 
O O O 



r-~ tr to 

O O rH 



u 


t--. 


CO 


a, 


CD 

>H 


rH 


r—l 




CO 




o 




m 


O 


•rH 


CN. 


Uh 


Ol 



CO 



r. 
CD 

X 

6 
3 
2 



C 

o 

CH .H 

O P 

3 
CD +H 
P-.-H 

>N+-> 

f-< in 

C 



o 
o 



o o o 
o o o 



o o o 
o o o 



o o to 
o a <y\ 



rH VO CN] 
rH O CNI 



vO 



vO vO In. 

|N~ t-~ IN 



|n- tO CTi 
r-~ vo O 



vO i-H 
O rH 



<* 


vO O O 


CNI LT1 rH 


CN] 


-r 


CNI 


H N IN 


tO CNI LO 


CNI 


to 


r-* 


r-t ^-i O 


rH rH rH 


r-H 


r-H 



o o 
o o 



to vO 

O i-H 



to 


cni co in 


LO CN] [N 


LO 


■^r 


to 


^3- to oo 


cni r- oo 


vO 


CNl 


o 


o o o 


o oo to 


o. 


to 



O CNI 

in o 



to 


*tf- CNI CN) 


Ol <3- LO 


vO Cn 


«3- 


LO tO CNI 


CO tO LO 


vo t-N 





in 


in 


O 




m <-< 


rH 


in O 




C o 


in O 


rH X 




O in O 


C in o 


o u 




■H rH x 


O rH X 


O CO 




+-> O u 


•H O U 


X 




3 O CO 


4-> O CO 


U rH 




4->X 


3 X 


CO CO 




•H CJ rH 


+J OH 


o 


in 


4-> CO CO 


•H CO CO 


i-{ -H 


c 


in U 


4J O 


co -d 


o 


Gr-< -H 


in rH -H 


U CD 


H 


h cO 13 


C CO 13 


•H 2 


+J 


U CD 


HOD 


13 


3 


<D-H 2 


•H 2 


0) 4-> 


4-> 


4-> T3 


U 13 


2 3 


H 


CO CD 4-> 


■H CD 4-> 


o 


4-> 


> s: 3 


rH 2 3 


X X 


in 


•H O 


X o 


4-> 4-> 


c 


Sh X X 


3 X X 


•H -H 


i — i 


Ph*JP 


Dh 4-> 4-> 


s s 



S H22 



<< 



-39- 



overall percentage gains between FY 1964 and 1974 were almost identical; 44 
per cent for all sources of R&D funds and 40 per cent for the federal compon- 
ent. Both showed marked gains in FY 19 7 3 over FY 1972, and both declined in 
FY 1974 (but federal R&D funds more sharply). 

Private vs. public universities . In overall percentage terms, private 

institutions showed an increase of 41 per cent in federal R&D funds 

over the 11-year period, as compared to 39 per cent for the public sector. 

The public institutions had a slightly higher increase after FY 1972: 

13 vs. 12 per cent for FY 1973, with a sharper drop in funding for 

FY 1974 to about the same level as the private sector (relative to the 

FY 1972 index base). 

Presence vs. absence of medical schools . Although the biomedical-behav- 
ioral funding trends are somewhat mixed, and different for private than 
for public universities, the R&D growth rates for institutions without 
medical schools tended overall to be somewhat higher than those for 
universities with medical schools. However, these composite trends 
mask rather clear-cut differences between private and public institutions: 
(a) for the private sector, institutions with medical schools show sub- 
stantially higher growth rates after FY 1972 than those without medical 
schools; (b) for public universities, the general trend was in the oppo- 
site direction. 

Again, it should be emphasized that in terms of actual expenditure levels 
in real dollars, institutions with medical schools, both public and pri- 
vate, stood far above those without medical schools (see Table 14) . 

Comparisons of Carnegie Commission categories of institutions . Institu- 
tional breakdowns of federal R&D funding by Carnegie Commission classes 
are shown in Appendix TABLES B-2 and B-2A. Comparisons of the means and 
index numbers in these tables should again be tempered by the fact that 
the funds for both private and public universities are heavily concen- 
trated in the Research Universities I category. 

Over the 11-year period, the following are the overall percentage changes 
in federal R&D funding: (a) Research Universities I (private, 51%; public, 
46%); Research Universities II (private, 7%; public, 19%); Other Categories 
(private, -14%; public, 48%). 

R&D Expenditures by Groups of Biomedical-Behavioral Disciplines 

The following sections will present detailed data for the four disciplines 

comprising the "biomedical-behavioral" complex as described above. Only those 



-40- 



for biological sciences, medical sciences, and psychology will be discussed 
in the text. Detailed tables are included in the text and in Appendix B for 
"Life Sciences not Elsewhere Classified"; but due to its miscellaneous nature 
and the relatively small funding levels, no discussion of this category has 
been included. 

Federally funded R&D expenditures for biological sciences . The means 
and index numbers for institutions classified by control and medical school 
status are shown in Tables 15 and 15A on the following page. The data for 
all institutions combined are similar in pattern to those shown in Tables 14 
and 14A for total biomedical-behavioral R&D funding, except that the biologi- 
cal sciences received a relatively much higher increase in FY 1973. (The 11- 
year increase for the total sample in federal R&D funding for biological 
sciences was 49 per cent.) 

Except for private universities without medical schools, all institutional 
subgroups participarted in the sharp funding increase in FY 1973 — with public 
institutions gaining 26 per cent and the private sector 17 per cent. But 
the latter suffered a relatively small drop in FY 1974, whereas the public 
universities lost almost all of their gain. Institutions without medical 
schools fared better than those with them in holding their post-1972 increases. 

The trends in biological R&D funding by Carnegie Commission categories 
are shown in Appendix TABLES B-3 and B-3A. There are striking differences 
between the private and public sectors in the post-1972 trends for the two 
classes of research universities. For Research Universities I, both private 
and public institutions had sharp FY 19 73 increases (22 and 27 per cent, 



-41- 



Table 15. Trends in Mean Federally Funded Expenditures for Bio- 
logical (Agriculture Included) Research by Type of Insti- 
tution — Constant Dollars in Thousands (NIH R&D Deflator, 
FY 1964 = 100) a 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 1974 



All Institutions 



143 



$1,105 $1,439 $1,527 $1,524 $1,860 $1,643 



All Private Institutions 54 
With Medical Schools 32 
Without Medical Schools 22 



1,107 1,599 1,540 1,704 

1,476 2,198 2,116 2,362 

571 729 701 746 



1,996 1,966 

2,865 2,797 

733 757 



All Public Institutions 89 
With Medical Schools 34 
Without Medical Schools 55 



1,103 1,342 1,519 1,414 

1,365 1,799 2,488 2,178 

941 1,059 920 942 



1,778 1,448 
2,638 1,910 
1,246 1,162 



All with Medical Schools 66 
All without Medical Schools 77 



1,419 1,992 2,308 2,267 2,748 2,340 
835 965 857 886 1,100 1,046 



Source: National Science Foundation. 



Table 15A . Trends in Index Numbers for the Means of Federally 

Funded Expenditures for Biological Research Shown in 
Table 15 (Means for FY 1972 = 100) 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 



1974 



All Institutions 



143 



72.5 94.4 100.2 100.0 122.1 107.9 



All Private Institutions 54 65.0 93.9 90.4 100.0 117.2 115.4 
With Medical Schools 32 62.5 93.0 89.6 100.0 121.3 118.4 
Without Medical Schools 22 76.5 97.6 93.9 100.0 98.2 101.5 



All Public Institutions 89 78.0 94.9 107.4 100.0 
With Medical Schools 34 62.7 82.6 114.2 100.0 
Without Medical Schools 55 99.9 112.4 97.6 100.0 



125.7 102.4 
121.1 87.7 
132.3 123.3 



All with Medical Schools 66 62.6 87.9 101.8 100.0 121.2 103.2 
All without Medical Schools 77 94.2 108.9 96.7 100.0 124.1 118.1 



-42- 



respectively) ; but in FY 1974, the private universities lost very little of 
their gain (3 per cent) while public Research Universities I lost about 29 
per cent. 

In the case of Research Universities II, the public institutions gained 
some 14 per cent in federal R&D funds for biology in FY 1973, with a slight 
increase in the following year. But the private universities in this category 
had a sharp decline of about 13 per cent in FY 197 3 and gained only 4 per cent 
from that level in FY 1974. 

Federally funded R&D expenditures for medical sciences . The means and 
index numbers for this variable are shown in Tables 16 and 16A on the following 
page. Comparison of the means for all institutions combined with those in 
Table 15 for biological sciences shows that those for medical sciences are 
somewhat higher. The trends in means for the two areas from FY 1964 to 1974 
are also different in pattern. The mean expenditure for medical sciences 
reached its peak in FY 1968; and, after fluctuating down and up, only managed 
to get back almost to the 1968 level in FY 1974. By contrast, the growth peak 
for biological sciences wasn't reached until FY 1973. Over the 11-year period, 
federal R&D funding for medical sciences increased 22 per cent vs. 49 per cent 
for biological sciences. 

The most striking difference shown in Table 16A between public and pri- 
vate universities in federal R&D funds for medical sciences is the opposite 
trends from FY 1973 to 1974. Private institutions show a decline of 8 per 
cent while the public group increases 21 per cent. 

Since virtually all of the federal R&D funds for medical sciences were 
granted to institutions with medical schools, trend comparisons of these 



-43- 



Table 16. Trends in Mean Federally Funded Expenditures for Medical 
Research by Type of Institution — Constant Dollars in 
Thousands (NIH R&D Deflator, FY 1964 = 100) a 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 1974 



All Institutions 



143 



$1,422 $1,749 $1,705 $1,581 $1,679 $1,742 



All Private Institutions 54 
With Medical Schools 32 
Without Medical Schools 22 



2,194 2,751 2,641 2,505 2,648 2,443 

3,651 4,569 4,441 4,209 4,431 4,095 

73 106 23 26 55 41 



All Public Institutions 89 
With Medical Schools 34 
Without Medical Schools 55 



953 1,142 1,138 1,021 1,092 1,317 

1,741 2,148 2,354 2,370 2,633 3,074 

466 520 386 187 139 231 



All with Medical Schools 66 
All without Medical Schools 77 



2,667 3,322 3,366 3,262 3,505 3,569 
354 401 282 141 115 176 



Source: National Science Foundation. 



Table 16A . Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Medical Research Shown in Table 16 (Means 
for FY 1972 = 100) 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 



1974 



All Institutions 



143 



89.9 110.6 107.8 100.0 106.2 110.2 



All Private Institutions 54 
With Medical Schools 32 
Without Medical Schools 22 



87.6 109.8 105.4 100.0 

86.8 108.6 105.5 100.0 

282.3 406.3 89.9 100.0 



105.7 97.5 
105.3 97.3 
210.1 155.9 



All Public Institutions 89 
With Medical Schools 34 
Without Medical Schools 55 



93.3 111.8 111.4 100.0 

73.5 90.7 99.3 100.0 

249.0 277.7 206.3 100.0 



106.9 129.0 

111.1 129.7 

74.0 123.3 



All with Medical Schools 66 81.8 101.9 103.2 100.0 107.5 109.4 
All without Medical Schools 77 250.7 284.4 200.1 100.0 81.2 125.0 



-44- 



universities with those without medical schools would be relatively meaning- 
less. 

The comparisons of federal R&D funding for medical sciences by Carnegie 
Commission categories are shown in Appendix TABLES B-4 and B-4A. These data 
show clearly that the decline in FY 1974 funding noted above for all private 
universities combined was concentrated in Research Universities I (which had 
most of the private R&D total for medical sciences) . By contrast, public 
Research Universities I had increases after FY 1972 markedly above those for 
the two other public categories. 

In percentage terms, the 11-year increase in federal R&D funds for 
private Research Universities I in medical sciences was 22 per cent, while the 
corresponding increase for public universities was 56 per cent. 

Federally funded R&D expenditures for life sciences not elsewhere classi- 
fied . Data for this category of disciplines are included in Tables 17 and 
17A (on the following page) and in Appendix TABLES B-5 and B-5A. But, as 
already noted, the small amounts involved and the miscellaneous nature of the 
disciplinary group do not make discussion of the results seem profitable. In- 
spection of the index trends in Table 17A shows how erratically the mean values 
fluctuate from year to year and from group to group. 



-45- 



Table 17 . Trends in Mean Federally Funded Expenditures for Research 
in Life Sciences Not Elsewhere Classified, by Type of 
Institution — Constant Dollars in Thousands (NIH R&D 
Deflator, FY 1964 = 100) a 



Type of 
Institution 



Number 



1968 



Fiscal Year 



1970 



1972 



1973 



1974 



All Institutions 



143 



$172 



$215 



$156 



$146 $148 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



54 


157 


156 


121 


203 


291 


32 


264 


264 


202 


311 


489 


22 


2 





3 


46 


1 


89 


181 


250 


177 


112 


63 


34 


178 


294 


286 


205 


104 


55 


182 


222 


107 


53 


37 


66 


219 


280 


246 


256 


288 


77 


131 


159 


77 


51 


27 



Source: National Science Foundation. 



Table 17A . Trends in Index Numbers for the Means of Federally Funded 
Expenditures for Research in Life Sciences Not Elsewhere 
Classified Shown in Table 17 (Means for FY 1972 = 100) 



Type of 
Institution 



Number 



1968 



Fiscal Year 



1970 



1972 



1973 



1974 



All Institutions 



143 110.2 137.9 100.0 



93.8 



95.2 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



54 


130.1 


129.3 


100.0 


167.9 


240.4 


32 


130.6 


130.4 


100.0 


153.6 


242.0 


22 


66.0 


9.6 


100.0 


1837.3 


49.5 


89 


102.1 


141.4 


100.0 


63.5 


35.8 


34 


62.2 


102.7 


100.0 


71.6 


36.5 


55 


169.7 


207.2 


100.0 


49.7 


34.5 


66 


89.0 


113.5 


100.0 


103.8 


117.1 


77 


168.8 


205.3 


100.0 


66.3 


34.6 



-46- 



Federally funded R&D expenditures for psychology . Comparisons of means 
and index-number trends are shown in Tables 18 and 18A on the following page — 
for groups classified by type of control and medical-school status. 

Federal R&D funds for psychology increased substantially (60 per cent) 
for all institutions combined between FY 1964 and 1968, but declined from an 
index level of 107.9 to 95.8 for FY 1974 — a six -year drop of 11 per cent. This 
trend was in sharp contrast to that for biological sciences which turned up- 
ward after FY 1972 (Tables 15 and 15A) . The general pattern for psychology 
resembled that for medical sciences up to FY 1973; but whereas federal R&D 
funds for the latter increased in FY 1974, those for psychology declined. 

Private and public institutions differed somewhat in patterns of change 
in federal funding for psychology between FY 1964 and 1972 — the differences 
being in the faster growth rate for public institutions and in a sharp drop 
for the private group in FY 1970. But between FY 1972 and 1974, the trends 
for both groups were alike: an increase for FY 1973 followed by a decline for 
FY 1974 (more marked for the private than for the public sector) . 

There was fairly close agreement between the trends for all institutions 
with and all without medical schools in federal funding for psychological 
research — as shown by the last two lines in Table 18A. 

The comparisons of federal R&D funding trends in psychology by Carnegie 
Commission categories of institutions are shown in Appendix TABLES B-6 and 
B-6A. As with the biological and medical sciences, the funds are largely 
concentrated in Research Universities I, which means that the relatively 
small average amounts in the other two categories (both public and private) 



-47- 



Table 18 . Trends in Mean Federally Funded Expenditures for Psychology 
by Type of Institution — Constant Dollars in Thousands (NIH 
R&D Deflator, FY 1964 = 100) a 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 1974 



All Institutions 



143 



$142 $227 $218 $210 $216 $202 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



54 


164 


204 


183 


193 


198 


178 


32 


233 


236 


219 


230 


235 


210 


22 


64 


157 


131 


138 


144 


132 


89 


129 


241 


239 


221 


227 


216 


34 


199 


417 


403 


371 


382 


366 


55 


86 


132 


138 


128 


132 


122 


66 


216 


329 


314 


303 


310 


291 


77 


80 


139 


136 


131 


135 


125 



% 



ational Science Foundation. 



Table 18A . Trends in Index Numbers for the Means of Federally Funded 

Research in Psychology Shown in Table 18 (Means for FY 1972 
= 100) 



Type of 
Institution 



Number 



Fiscal Year 



1964 1968 1970 1972 1973 



1974 



All Institutions 



143 



67.7 107.9 103.6 100.0 102, 



95.8 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



54 


85.2 


105.6 


95.0 


100.0 


102.6 


92.6 


32 


101.4 


102.5 


95.1 


100.0 


102.1 


91.4 


22 


46.1 


113.3 


94.7 


100.0 


103.8 


95.3 


89 


58.5 


109.0 


108.2 


100.0 


102.9 


97.5 


34 


53.7 


112.3 


108.8 


100.0 


102.9 


98.7 


55 


66.9 


103.1 


107.2 


100.0 


102.9 


95.4 


66 


71.3 


108.7 


103.7 


100.0 


102.6 


96.0 


77 


60.7 


106.2 


103.4 


100.0 


103.2 


95.3 



-48- 



tend to fluctuate more erratically from year to year than that for the top 
category. Private Research Universities I had a relatively high level of 
federal funds for psychology R&D in 1964 as compared with public institutions 
in this category (index numbers of 84.6 and 58.1, respectively, in terms of 
the FY 1972 base). But the public group grew rapidly through FY 1970 as com- 
pared with the private sector (index of 114.2 vs. 92.4 for that year). There- 
after their trends followed fairly similar patterns through FY 1974, with the 
private group declining to a lower level than the public (91.3 vs. 98.9). 



-49- 



VI. TRENDS IN UNIVERSITY ENROLLMENTS 

The enrollment statistics used in this study came from two REGIS surveys 
conducted by the National Center for Education Statistics: (a) the survey of 
opening fall enrollment, which was limited primarily to numbers of students 
by educational level, sex, and full-time or part-time status; (b) the survey 
of enrollment for advanced degrees by academic discipline. 

Trends in the following categories of enrollment have been analyzed by 
type of institution: total degree-credit enrollment; enrollment for advanced 
degrees in all fields; and enrollment for advanced degrees in biomedical- 
behavioral sciences — all fields combined, together with breakdowns into the 
three component fields included in this HEGIS survey: biological sciences, 
health professions (but not M.D.'s, D.D.S.'s, etc.), and psychology. It 
should be noted that the enrollment category "biomedical-behavioral" differs 
in composition from the NSF expenditure category of the same name in that 
"agriculture" is included in the latter composite but not in the former. This 
"mismatch" should be taken into account in evaluating comparisons between 
enrollment and financial trends. (The NSF category "life sciences not else- 
where classified" is also missing from the HEGIS classification of disciplines 
but these fields presumably are included among the "biological sciences.") 

Following the plan used in the analyses of financial data, results are 
presented for two sets of institutional classification: (a) the subdivision 
of private and public universities by medical-school status; (b) the subdivi- 
sion by categories of the Carnegie Commission on Higher Education. The tables 
of data for the first type of classification have been inserted into the text as 



-50- 



they are discussed, while those for the Carnegie Commission classification have 
been grouped into Appendix C. As was done in the preceding chapter on biomedi- 
cal-behavioral expenditures, the discussion of the results for both types of 
breakdowns will be integrated by enrollment category. 
Total Degree-Credit Enrollment 

The variable used as a measure of total fall-term enrollment was the 
"head count" of all students registered for courses creditable towards a degree. 
Both on-campus and extension students were included, but not students enrolled 
for non-degree-credit courses. No distinction was made between full-time and 
part-time students. 

Usable data were available for only eight of the 11 years covered in the 

present study — the omitted fiscal years being 1964, 1967, and 1968. The mean 

number of enrollments per institution is shown in Table 19 on the following 

page for the 147 institutions that supplied usable data — with breakdowns in 

terms of medical-school status. Comparative trends are shown in Table 19A, 

which presents index numbers for the means shown in Table 19 (with the mean 

for FY 1972 as the base) . 

All institutions combined . Total enrollment increased rapidly between 
FY 1965 and FY 1971 (from an average of 11,568 to 16,474), but levelled 
off thereafter through FY 1974. A comparison between the growth pattern 
for total enrollment and the trend in educational-and-general revenues 
is shown in the following index-number series for the two sets of means: 

1969 1970 1971 1972 1973 1974 

Degree-credit enrollment 90.3 94.7 100.6 100.0 102.0 102.6 
Total E&G revenues 92.1 95.7 97.7 100.0 100.0 102.3 

Obviously, the relative growth rates for enrollment and E&G revenues over 
the six-year period are quite similar. But, as was noted earlier, the 
percentage increase in enrollment was slightly greater than that for E&G 
revenues (13.6 vs. 11.1 per cent). 



-51- 



to 
t-- 
01 



oo 1001 to i~~ \o m m *a- 

O mo O 00 Ol cm *»• o 

oo vol- vo i- ■* o cm vO 

vo oh oi o »*■ oo oo m 

rH rHlH rg N H rH rH 



O0 vOtO 00 O i-H t— "It-- 

O vOtO CI NNH to Ol 

r- lot to to o o o m 



I— 1 


OrH CI 

rHrH 


O «* OO 
CM (VI i-l 


oo m 


to 


i-hcm r-. 

CIO i-l 
I/IN O 


HOOH 
to CM v© 

oo lo m 


oo r» 
O <vl 


o 


OH Ol 


oint- 


t— m 






<0o 
uh 



Oo 
r— 
►, r ci 

J-JiH 
HI 

H 



U-.VO 
CT1 



vO 
vO 
Ol 



in 
vo 

Ol 



«* I-.CV1 o"> 00 i-l o to o 

r-» oo at «» o cm rH o 
»* oo m r-~ t-~ to i-i rH 



cm ocm \o riioo m r-~ 

O rHrH r-~ NOON tO rH 

in vom to <t oh o cm 



O tor— CM O 'S' ^ 

oo torn o <vi oo in 
r-~ in»t to to vo cm 



T 
•* 



CO* N\00 (MOO 
*J-m vO vO 00 CM r- o 

oot-~ in ih *» h cm o 



00 rHt-. in rH '5* O rH 00 

\o cmcticm mincM oo 
in "»* oi oo vo i— i i—i cm 



<D 
X> 

e 

3 

z 



mcvi to 
into cm 



C 

o 

IH -H 
O 4-> 
3 
0) 4-> 
CVH 
>v+J 

H m 

c 





c o 




O in O 




•HrH J3 




*-> o u 




3 O Ui 




*JX 




•HUH 


W 


4->co « 


C 


in o 


o 


CrH -H 


H 


i— t rt t3 


«-> 


o <a 


3 


oi-h s 


<-> 


+J-T3 


H 


CO 0) <-> 


+J 


>2 3 


in 


•H O 


c 


I-.C JS 


hH 


0»4-> 4-1 




•H -H 


rH 


H3S 



o to in 



in 

rH 

in o 
C in o 

O rH JT, 
•H O U 

4-> o to 

3* 
(J OH 
•Hlfl « 

4-> o 

in rH -H 

C rt-O 

H O 4) 

■hS 

•H d> 4-» 
rH 2 3 
rO O 

3rCrS 

CU 4-> +J 

•H -H 

HSS 



r~- o 

vO OO 



in 

rH 

o 
in o 
>-h,c 
O U 
o to 

O rH 

u 

rH -H 

a) -a 
u a> 
•h 2 

-a 

oi +-> 

2 3 

o 

rCrC 
4-> 4-> 
•H -H 

» 3 



<< 



in 



4-> 
CO 



c 
o 



u 

3 

•a 

Ih 
O 

4-1 

U 

a> 
+j 

c 
<u 
u 



cd 
C 
O 



C8 



u 

Ih 

3 
O 

to 
at 



o 

r-l 
Xi 
CO 
f- 

c 



o 

XI 

to 



c 

01 

E 



o 
u 

a 
w 



•3 

a> 

Ih 
U 

i 

0) 
<0 

u 

00 

a) 
Q 

m 
o 

in 

c 

CD 
V 

s 

01 

rC 
+J 

u 
o 

>4H ^ 

o 

in o 

U rH 
0) 
XI II 

e 

3CM 

a> 

X rH 
0) 

•§£ 

u 

c o 

•H >M 

in in 

T3 C 

C CO 

01 01 

r<2 

E-.V-/ 



cs 

ai 



o 



Ol 

vO 

Ol 



LO 

Ol 



01 
XI 



C 
O 

m -h 
o +j 

3 
V +-> 

CVH 

H in 
C 



CV1 OHOl tO**CM tOCM 

O OOOl ooo oo 



CM OrHOl CMCMCM (V|(V1 

O OOOl ooo oo 



o 
o 



o 

o 



Ol 



ooo 
ooo 



ooo 
ooo 



o o 
o o 



IO vO Ol 
O O Ol 



Ol o oo 
Ol O Ol 



to oi in oi oi rH 



o i— i r- 

O O Ol 



CM to CM 

Ol Ol Ol 



Ol tO rH ■>* tO O 






o 



cm in oi 

oi oi oo 



rH «3- Ol 



Ol tO CM 
CO Ol oo 



Tt vO to 
vO vO O 



in 

c 
o 



in 

c 



in 

U) rH 

a o 

O in O 

•hhj: 

4-" O O 

3 O to 

•H U rH 

•M to n) 
in u 

CHH 
hH C3 13 

U 01 

oi -h 2 

+J rg 
(D 01 ♦-) 
>2 3 

•H O 

t- x: xi 
a< +j +j 



in 

1— I 
in o 
C in o 

OrH,C 
■H O U 
4-> O tO 
3rC 

+J U rH 

■h to rt 

4-1 O 

in •-< -h 

C CTJ TD 

H O 01 

•H 2 

O TJ 
•H 01 4-> 
rH2 3 

Xl o 
3hCC 

Oh 4-> 4-> 



CM O0 

O Ol 



lO to 
Ol Ol 



to 


<tHlv 


to Ol Ol 


Ol vO 


o 

Ol 


OlHvO 
Ol O Ol 


f-. Cv. \0 

OO 00 oo 


rH 00 
Ol OO 



O CM 
00 t— 



■*f VO 

r~- vo 



[~~ 


in cm to 


NlflN 


r- o 


**■ 


in to cm 


oi to in 


vo OO 



rH, 

o 
in o 

rH x: 

o o 
o to 

rC 

U rH 
CO CD 

u 

rH -H 

a) -a 

U <ti 
•H 2 
T3 

01 4-> 

2 3 
o 

4-> 4J 

■H -H 



rH HJS HJg rH, 



< < 



-52- 



99.4 


100.3 


103.8 


100.0 


100.7 


100.6 


95.1 


95.7 


96.3 


100.0 


96.4 


96.9 


87.3 


92.9 


99.6 


100.0 


102.5 


103.3 


89.5 


95.7 


99.0 


100.0 


103.2 


107.1 



Private vs. public institutions . Inspection of the trends for private 
and public institutions in Table 19A shows that the respective patterns 
differ significantly. The index numbers for all private institutions 
range from 89.0 in FY 1965 to 100.6 in FY 1974, whereas the corresponding 
indices for public institutions are 65.8 and 103.3. Moreover, the private 
institutions as a group had substantially reached a plateau by FY 1969 
that was maintained at an index-level of about 100.0 until FY 1974 (except 
for a one-year increase to 103.8 in FY 1971). 

It will be recalled from chapter III that the private and public trends 
in E&G revenue growth also differed significantly from FY 1969 through 
FY 1974. A comparative summary of the enrollment and revenue means for 
the two sectors is presented in the following tabulations of index num- 
bers for the two types of data (from Tables 19-A and 5-A, respectively) : 

1969 1970 1971 1972 1973 1974 

Private Universities 

Degree-credit enrollment 
Total E&G revenues 

Public Universities 

Degree-credit enrollment 
Total E&G revenues 

Obviously from these figures, the E&G revenue trend for private institu- 
tions declined relative to enrollment level after FY 1972, whereas for 
public institutions the constant-dollar growth in revenue exceeded the 
rate of overall enrollment increase. 

Presence vs. absence of medical schools . The degree-credit enrollment 
trend data in Table 19A do not show marked differences between universi- 
ties with and those without medical schools — whether comparisons are made 
within the private sector, within the public sector, or for the two sec- 
tors combined. 

Enrollment differences among Carnegie Commission categories of institutions , 
Data for the classification of the 147 institutions by the Carnegie Commi- 
ssion categories are shown in Appendix TABLES C-l and C-1A. The compara- 
tive enrollments of the private and public institutions in these cate- 
gories are shown in TABLE C-l — over the period from FY 1965 through 1974. 

The comparative trend indices in TABLE C-1A show interesting differences 
among the private institutions in comparative growth rates. Research 
Universities I have index values of 98.0 for FY 1965 and 101.0 for FY 
1974, showing little overall change. But during the three-year period 
1969-1971 their enrollments rose appreciably above this plateau, as 
shown by the respective index numbers of 105.2, 106.2, and 109.5. The 
other two categories of private institutions began at levels considerably 
lower than Research Universities I, and showed fairly steady growth rates 
overall to FY 1971, levelling off thereafter. 



-53- 



Among the public institutions, the enrollment trends for the three Carne- 
gie Commission categories had essentially similar patterns of growth 
from FY 1965 through FY 1974, with Research Universities I showing some- 
what higher increases in FY 1971 and 1974. 

Enrollment for Advanced Degrees in All Fields 

Data on enrollment for advanced degrees by disciplines were available for 
only 121 institutions over the seven-year period (FY 1967-1973). Unfortunately, 
the computer tape with FY 1974 data had not been released in time for use in 
this study. 

The mean numbers of enrollments for advanced degrees per institution for 

all graduate fields are shown in Tables 20 and 20A on the following page. 

All institutions combined . The trend indices in Table 20A show that 
graduate enrollment increased from an index number of 79.2 in FY 1967 
to 100.7 in FY 1973 — an overall increase of 27 per cent. But there was 
essentially no growth after FY 1971 when the index number was 100.5. 

Comparisons of the growth trend for advanced-degree enrollment with those 
shown earlier for all sponsored R&D revenues and for federally sponsored 
R&D revenues are made in the following figures: 

1969 1970 1971 1972 1973 



Advanced-degree enrollment 91.6 96.9 100.5 100.0 100.7 
All R&D revenues 102.1 101.6 96.8 100.0 99.9 

Federal R&D revenues 106.2 104.2 96.4 100.0 98.7 

The results of these comparisons may be summarized as follows: (a) over 
the five year period, enrollment for advanced degrees increased 10 per 
cent; (b) all R&D revenues decreased two per cent; (c) federally spon- 
sored R&D revenues decreased 7.6 per cent. While the differences be- 
tween the enrollment and revenue trends might seem to be small in magni- 
tude, they probably were not negligible in impact upon the graduate 
departments of universities. Furthermore, the divergencies would very 
likely increase if data were available to carry the comparisons through 
the fiscal years 1974, 1975, and 1976. 

Frivate vs. public institutions . Private institutions maintained a more 
nearly constant level of enrollment over the entire seven-year period 
than did the public institutions. Table 20A shows index changes from 
87.9 to 101.0 for the former and from 73.9 to 100.5 for the latter. 
Moreover, private institutions had reached a plateau by FY 1970 while 
this occurred for the public institutions in FY 1971. 



-54- 



i 








*» 


HlOO 


to cm r~ 


O rH 




















c 




IO 


00 


OOICO 


MW^ 


00 CM 








to 


r- 


o m «*• 


m to o 


00 H 


o 




r^ 


o 


OOIOO) 


CM H O 


r- ** 








H- 


• 








•H 




en 


» 














cn 


o 


rH O CM 


o en cm 


en cm 


4-> 




rH 


to 


CM tO H 


to 'a- cm 


IO CM 




rH 




rH 


o 


O O O 


o en o 


en o 


3 
















rH 






rH 


rH rH rH 


rH rH 


H 


♦J 

•H 
















< 














*J 
















a 














in 






evi 


PJ O tO 


LO rH tO 


O0 rH 




■H 














C 




CM 


\o 


IvOOM 


>© oo cn 


oo r~ 








CM 


O 


O O O 


o o o 


o o 


hH 




r- 


O 


SKlOl 


CM rH lO 


l"~ to 




in 




f-. 


a 












Ol 


•• 










4) 




en 


O 


o o o 


o o o 


o o 


4-1 




rH 


IO 


cm to rH 


tO f CM 


tO CM 




4) 




rH 


O 


O O O 


o o o 


o o 


O 


U 














rH 

oo 


H 




rH 


rH rH rH 


rH rH rH 


H H 


4) 


c8 














41 


CO 












P. 


4> 














a 


4> 












>"« 


>« 




vO 


m cm o 


«9- cn CM 


00 rH 






>• 












H 




rH 


r~ 


OlOlN 


r- oo to 


«9- T* 




-o 




rH 


in 


oo to en 


IO o o 


>o 00 




rH 


r- 


o 


MOOl 


CM CM lO 


oo to 




CO 


rH 


r-. 


> 








>s 


cd 


en 


•» 










u 


cr) 


en 


o 


O © rH 


O CM p» 


H CO 


.a 


(J 
in 


rH 


IO 


cm to rH 


to "S- CM 


tO CM 




e 

CO 


u 

in 


rH 


o 

rH 


o o o 

rH rH rH 


o o en 

rH rH 


o en 

H 


W 


•H 














> 


•|H 












■a 


u. 














•d 


0. 












|H 
















< 














<U 


m 




vO 


*WH 


Ol CM IO 


«*• >o 






<4H 












• H 


o 


o 


vO 


t-.NOO 


cn cn r~ 


vO o 




rH 


o 


o 


en 


rH tO tO 


en en ui 


"* H 


w. 




r-. 


CTl 


r- nf 00 


o o to 


r- cm 




O 




p- 


• 










e 


0\ 


* 










<4h 


a 


<& 


o 


OHS 


H» t-~ rH 


en to 


H 


Hi 


rH 


CM 


NMH 


to «a- cm 


tO CM 






Ih 


rH 


en 


o o en 


en en en 


en en 


H 


CO 














•l-> 


4) 






rH rH 






< 


H 














4) O 


H 












c 


H 














e o 


rH 












•H 


rH 




to 


rH tO m 


m m r~ 


tO rH 




rH rH 


rH 














ed 


CTl 


o 


O ^- rH 


NH00 


00 -O 




rH 


CO 


en 


vO 


•rfClOl 


rH cm to 


>o en 


VI 


PL, 


NO 


00 


MOM 


00 O0 rH 


LO O 




O 11 


Pm 


VD 


* 








0) 




Cn 


«> 










u 




en 


rH 


r- oo to 


00 rH ^J- 


«t o 


a> 




rH 


<vl 


CM tO H 


CM IO CM 


tO CM 




CCM 




rH 


CTv 


CT* 0> O 


oo en oo 


en oo 


H 
















W r~- 














00 
















en 














4> 
















<H rH 














Q 






oo 


t~ to to 


CM O r*- 


en »»• 




o 














■a 




SO 


r>- 


00O G) 


toto 


to IO 




in u. 




90 


m 


to \o r>- 


o o m 


00 ^ 


CD 




<o 


v© 


WHN 


r~ r~ © 


«* en 


• 


a 




o 


• 








U 




Cn 


* 








in 


CO H 




en 


r~ 


to to CM 


«* oo oo 


O CM 


c 




rH 


CM 


CM to H 


CM to CM 


IO rH 


O 


4) O 




rH 


00 


en en en 


oo oo r» 


en oo 


(« 














•H 


S4H 














> 














+J 
















•a 














in 


4) tn 














< 






T 


00 vO to 


MNt 


«* r^ 


•H 


rC C 

4-> CO 














l-c 




r~ 


CM 


tOlflN 


rH oo r^ 


cm m 


a) 


4) 




r-. 


CM 


en m rH 


en •£> <«■ 


m h 


O 




>o 


"» 


"» or-. 


<tNN 


rH f~ 


•4-> 


hS 




« 


a 








■4-1 




Cn 


* 


* » » 


* « *. 


M M 


CO 


O^-' 




en 


en 


K t-. en 


to oo oo 


CM •* 






rH 


CM 


NN H 


CM to rH 


IO tH 




tH 




rH 


r» 


oo oo oo 


t~ t~ \0 


oo r-» 


♦J 














c 


O 














c 














o 


in cm 














V 












•H 


u 












e 












■M 


CD 4) 












■— i 












CO 


rOrH 












H 


Ih 










u 


6 rd 


HI 










o 


CO 










3 


3 a) 


CO 










(h 


43 


rH 


o en rH 


HOH 


en cm 


TJ 


Z H 


43 


rH 


OOIH 


HOH 


en cm 


c 


s 


cm 


tO CM CM 


r-~ to <* 


in vo 


W 




e 


CM 


m cm cm 


t~. to ** 


m \o 


PJ 


Z 


rH 








M 


4> -H 


3 

z 


rH 








a 












O 


•a 












a) 










in 


MH 


a s 










m 


<t> 










rH 




HH * 










H 


2 






in 


in 


o 


H 


o 






in 


in 


O 








V) H 


rH 


in o 


0) 


(3 43 






in rH 


H 


m o 


(3 






c o 


in o 


I-H..C 


+-> 


H tO 






C O 


in o 


H J. 


• H 






O in O 


e in o 


o o 


c 








O in O 


C in O 


o u 








•HrH,C 


OHX 


O 00 


CD 


in in 






•H rH JC 


OH j- 


o to 


in 






♦J o u 


•H O U 


J2 


CJ 


TJ-O 






•M O O 


•H O O 


43 


TJ 






3 O CO 


«-> o to 


O rH 




C <H 






3 O to 


+-> o to 


O H 


c 


C 




•!-> 43 


3rC 


CO ed 


rH 


4> 4) 


c 




♦-43 


343 


00 CD 


a> 


o 




•HOH 


r> UH 


o 


at 


Ih .H 


o 




■H O rH 


•H> O H 


u 


u 


"4-1 -H 


CO 


-H> 10 CD 


•H tO CS 


rH -H 


C 


H U, 


<4H -H 


m 


■w to <a 


•H to W 


H -H 


E- 


O P 


C 


in o 


<-> U 


ee *a 


O 




O -P 


C 


in u 


+J o 


18 TJ 




s 


o 


e ih -h 


Ifl rH -H 


U V 


•H 




3 


o 


C rH-H 


in H -H 


O 4) 




4> *J 


•H 


H C8 TJ 


C cSTJ 


•HE 


+J 


a 


CO +-> 


•H 


HH C8 TJ 


C COT) 


•HE 


• 


tVH 


+J 


O 4) 


HH O 4) 


-0 


CD 


< 




C-'H 


+J 


U 4> 


hh O CD 


TJ 


o 




x*-> 


3 


fl)-H 2 


•H S 


0) +J 


Z 


o 




X+J 


3 


41-H S 


•H S 


4> H> 


rj 




H w 


+J 


«j*a 


OTJ 


E 3 




CM 




H m 


4-> 


•H> TJ 


O TJ 


E 3 






C 


•|H 


(8 4> *J 


•H 4) -H> 


o 








e 


•H 


C3 1) (J 


■H CO 4-> 


O 


CD 




t— i 


+J 


>S 3 


HE 3 


43 rC 


■ ■ 


co 




HH 


+J 


> S 3 


■HE 3 


X43 


rH 






(0 


•H O 


rO O 


4-> <M 


a> 


rH 






in 


• H O 


43 O 


■M H> 


43 






c 


(h^-C 


3J=£ 


•H -H 


u 


43 






e 


h.43 43 


343 43 


•H -H 


rt 






H 


Cm *-> H> 


Cm 4-> 4-> 


3 3 


f» 


0) 






HH 


Cm h> *-> 


&. 4-> <-> 


3 3 


E- 








•H .H 


•H -H 




3 


H 








•H -H 


•H -H 








rH 


tHS S 


rHS S 


i-t rH 


O 






rH 


HS?S 


HSS 


H H 






rH 


rH 


rH 


rH rH 


co 






rH 


rH 


H 


H H 










< 


< 


< 


<< 


crj 










■H, 


< 


< 


<< 



-55- 



There were no clear-cut or consistent relationships between R&D funding 
trends and trends in advanced-degree enrollment, for either private or 
public institutions. It may be recalled that all R&D revenues combined 
for the private sample tended to decline progressively from FY 1969 
through 1973, whereas the opposite occurred for the public institutions 
(see Table 9A) . In percentage terms, the mean of all R&D revenues 
declined about 9 per cent for private institutions and increased 10 
per cent for public institutions, from FY 1969 through 1973. For fed- 
erally sponsored R&D revenues, the percentage decline for private uni- 
versities was about 14 per cent, while the public universities gained 
some 7 per cent over the same period (see Table 11A) . 

Presence vs. absence of medical schools . The figures in Table 20A show 
only slight differences in advanced-degree enrollment between trends for 
private institutions with and those without medical schools. In the 
case of public institutions, those without medical schools showed a 
relatively greater growth rate over the seven-year period than did insti- 
tutions with medical schools. The former's index numbers ranged from 
68.4 in FY 1967 to 102.0 in FY 1973 while for public universities with 
medical schools the corresponding range was from 78.6 to 99.3. The 
respective percentage increases were 49 and 26 per cent. 

Advanced-degree enrollment by Carnegie Commission categories of institu- 
tions . Trend comparisons are shown in Appendix TABLES C-2 and C-2A 
for private and public institutions classified in the three Carnegie 
Commission groups. Research Universities showed the least amount of 
increase over the seven years, for both private and public universities 
(9 and 20 per cent, respectively). 

The greatest growth rate occurred for "Other Categories" of public insti- 
tutions, whose range in index numbers was from 59.2 in FY 1967 to 103.1 
in FY 1973 — an increase of 74 per cent. The corresponding gain for pri- 
vate universities was 22 per cent — a gain identical to that for private 
Research Universities II. For public Research Universities II, the 
seven-year increase was 44 per cent. 

Enrollment for Advanced Degrees in Biomedical -Behavioral Sciences 

As indicated at the beginning of this chapter, trend data will be pre- 
sented separately for advanced-degree enrollment in the following classes of 
biomedical-behavioral disciplines: all biomedical -behavioral fields combined; 
biological sciences; health professions; psychology. 



-56- 



All biomedical-behavioral fields combined . The total number of institu- 
tions with records of enrollment for advanced degrees in biomedical-behavioral 
sciences from FY 1967 through FY 1973 was 119. The mean enrollments per insti- 
tution over that period are shown in Table 21 on the following page, together 
with the corresponding index numbers for these means in Table 21A. 

All institutions combined . There was a progressive increase for the aggre- 
gate of all institutions from an index value of 75.5 in FY 1967 to 105.2 
in FY 1973 — an increase of 39 per cent. The rate of increase for bio- 
medical-behavioral sciences was thus considerably higher than that for 
all graduate fields combined (27 per cent) . 

The index numbers for advanced-degree enrollment may be compared with 
indices for mean R&D expenditures in biomedical-behavioral sciences 
(total and federally funded) for the years with data for both variables. 
The comparative figures for all biomedical-behavioral fields combined 
are as follows: 

1968 1970 1972 1973 

Enrollment for advanced degrees 
Total R&D expenditures 
Federal R&D expenditures 

The overall percentage increases from FY 1968 to FY 1973 for these three 
trend series are as follows: enrollment for advanced degrees, 30 per 
cent; total R&D expenditures, 16 per cent; federally funded R&D expen- 
ditures, 9 per cent. 

Private vs. public institutions . The index trend for all private insti- 
tutions shown in Table 21A parallels fairly closely that for all insti- 
tutions combined — the increase from 82.8 in FY 1967 to 108.7 in FY 1973 
being 32 per cent (as compared with 39 per cent for all institutions 
combined) . 

For all public institutions, the growth rate was higher than that for 
the private sector (43 versus 32 per cent) . But the private universities 
increased 8.7 per cent from FY 1972 to 1973, in contrast to the public 
increase of 3.6 per cent. 

Presence vs. absence of medical schools . The mean indices for combined 
private-public enrollment for advanced degrees shown in the last two 
lines of Table 21A indicate a greater gain for universities with than 
for those without medical schools (43 vs. 33 per cent). 



80.7 


89.1 


100.0 


105.2 


95.8 


100.4 


100.0 


110.6 


103.3 


105.6 


100.0 


112.4 



-57- 



IO 

[>- 
01 



Ol 



u 
CO 
4) 
>- 



o 

in 

•H 

tL, 

<4-l 

O 

e 

Si 
4> 
(- 



oi 



o 
en 



f- o> oo vo to to oo 

to cm to i-i <toN 



>0 t-- \0 Ol OINH to vO 

i/i \omt Hoto ooio 
to nmh *t no cm «rj- cm 



1/1 UllflN CM 00 l/l U1H 

CM tNrO 00M\O MN 

tO NIOH tO l/> CM *f CM 



p~ o «» cm cm i-^ cm oil— i 

iH t H^f t~-HlO r- I CM 

tO NMH (OWN "*• CM 



On 



f 


*»• r~ oo 


to "3- l/l 


oi m 


o 


to o to 


lo oo i/i 


Ol iH 


to 


cm to .-t 


IO «* CM 


tO CM 



oo 



o 

Ol 



oo ihcmlo oor^oo r*^to 
oo tooto cm rt- to r*» o 

CM CMIOrH tOfCM IOCM 



oi i— * r>- o toooo oito 

vO CMOICM OHH U100 

CM CM CM rH tO **■ CM Mr I 



u 
<u 

XI 

e 

3 

z 



c 
o 

<+H -H 

O P 

3 

4> P 

CVH 
XP 

E-> m 

C 



in 
C 
O 



3 
P 



ol 

C 



01 00 rH 
Tt CM CM 



in 

V) r-i 

e o 

o mo 

•H rH,C 
•POO 
SOW 

P X! 

•H UH 

P to n) 

Ul U 

C rH -H 

I— I CS Tj 

U 4> 

0).hS 

Cfl CD +J 

>Z 3 

•H O 

-h x: x 
o, <-> ■*-> 

•H -H 



O O O 

MOV 



in 

iH 

in o 
C m o 

ohx 

•H O O 
P O CO 

3 X 

P O i-l 
•H CO CO 

P O 

in rH -h 
C ei)T3 
(-1 O 4) 

■H S 

O 13 
•HDP 

HJ3 

X> O 

3xx 
a, p +j 



00 i-l 
l/l O 



in 

r-i 

o 

in o 

•H X 

o u 

O CO 

X 

U r-l 
CO C3 

u 

rH -H 

td T3 
u 4> 

•HS 

•a 

S 3 

o 

xx 



* * 



HJJ iH rH 

■H i-l r-l 

< << 



«1 
U 



co 
p 

CO 



o 



B) 

u 
3 
13 
CU 

iH 
O 

<P 

u 

4) 

p 

c 
a> 
u 



c 
o 



a) 
Z 



4> 

u 
ll 

3 

o 

CO 

R) 



u 


■H 


T> 


4) 


s 


o 


•H 


co. 


c 


•H 


m 


V 


a) 


(H 


oo 


<!>,—> 


Q O 


o 


TJH 


a> 


u n 


c 


a) cm 


> r- 


13 oi 


<-H 


rH>« 


O CL, 


CM 


Ih 


P o 


C<P 


4) 


e w 


-H c 


i-H n) 


O 4> 


n2 


C w 


w 


1— ( 


*+■! (NJ 


o 


V 


in iH 


CX> 


c« rt 


v h 


S 


c 


>- -H 


o 


<p p 


in o 


iix: 


4) CO 


X 


e i/> 


3 0) 


Z O 


c 


X 4) 


4) -H 


T3 y 


C CO 


i—i 


rH 


C R) 


•H i- 


O 


«1 -H 


13 > 


C CO 


«£ 


fH 4> 


h m 



1= 






ai 



o 



C3^ 



GO 






m oo oo rH 
o o O iH 



to i/i o 
o o o 



vO to 

o o 



o 
o 



o o o 
o o o 



o oo 
o oo 



o o 
o o 



fH 1H1HCM iHOl^J" OtO 

Ol OlOlOl OlOOOl OlOl 



Ol 
OO 



to 
oo 



oi oo i/i 

oo oo oi 



oo i/i to 

00 00 Ol 



vO iH vO Mf lO 



r-^ vo cm 

00 00 Ol 



•*t O O 

oo oo oi 



4> 

X 

e 

3 

z 



e 
o 

IH-H 
O P 

3 

1) p 

D.-H 

>^p 

H in 

B 



c 
o 



3 
p 



in 

a 



in 
in iH 
c o 
o in o 

•H iH J5 

P o u 
3 O CO 

P JS 

•H U rH 
P CO IS 
in u 

C iH -H 

►h a) 13 
U 4) 
4)-H S 
P 13 

ca 4) p 
>S 3 

•H O 

&< p p 

•H -H 

■hs a 

■H 

< 



in 

rH 

in o 

C in o 

OH£ 
•H O O 
P O CO 
3 J3 

P O rH 

■h co n) 
P u 

m ih -h 
a m i3 

l-H O 4) 

•HS 

U 13 
•H 4> P 

rH 2 3 

X) o 
3 J3 J5 
&, p p 

•H -H 



vO tO 
OO Ol 



"* o 

CM rH 

00 Ol 



r-~ 


"tOOl 


NNIs 


O iH 


o 

oo 


\D 1/1 O 
00 O0 Ol 


00 ■* T3- 

r~- r^ oo 


00 \o 
t^ oo 



l/> 


00 -3- Ol 


CM V 00 


■^- r-- 




CM tO O 
O0 00 00 


M OHO 

r~ >o r~- 


i~- t-. 



Ol 


Ol 00 rH 


O o o 


00 rH 


iH 


»*■ CM CM 


MO"t 


1/1 vO 



in 

iH 
O 

in o 

■H X 

o u 

O CO 

JS 

O fH 

CO n} 
O 

rH -H 
CO 13 
U 4> 

•H 2 

T) 
4) P 

S 3 
O 

XIX 



-* 5 



<< 



-58- 



But when similar percentages were calculated for the gains of private 
and public institutions separately, those for the public universities 
conformed to the pattern just described for the composite groups whereas 
the private universities showed a reverse trend. 

Advanced-degree enrollment in all biomedical-behavioral fields by 
Carnegie Commission categories of institutions . These enrollment-trend 
data are shown in Appendix TABLES C-3 and C-3A. The trend index patterns 
for all of the Carnegie Commission categories of institutions combined — 
private and public — were generally quite similar: progressive growth in 
total advanced-degree enrollment from FY 1967 through FY 1973. 

But there were significant differences in growth rates among the Carnegie 
Commission categories, which were similar for private and public univer- 
sities: the "Other Categories" of institutions had considerably higher 
growth rates than Research Universities I and II. For private institu- 
tions, the latter two groups showed seven-year increases of 29 and 30 
per cent, respectively, while the increase for "Other Categories" was 
38 per cent. For the public institutions, the corresponding figures 
were 39 per cent for Research Universities I and II, and 65 per cent 
for "Other Categories." 

Advanced-degree enrollment in biological sciences . The enrollment-trend 

data for graduate biological sciences are shown in Table 22 and 22A on the 

following page, and in Appendix TABLES C-4 and C-4A. 

All institutions combined . Enrollment grew at a slow but fairly steady 
pace from FY 1967 through FY 1973. The overall increase was 16 per cent. 

The following is a comparison of advanced-degree enrollment trends with 
those for R&D expenditures in biological sciences, for the years common 
to the two variables: 

1968 1970 1972 1973 

Advanced-degree enrollment 
Federal R&D expenditures 

Private vs. public universities . Private institutions showed only slight 
growth in enrollment over the seven-year period (7 per cent), while the 
increase was 20 per cent for the public group. 

Presence vs. absence of medical schools . The trend differences were slight 
between the composites of these two categories of institutions, with a 
single exception: from FY 1972 to 1973, institutions with medical schools 
increased four per cent in graduate biological-sciences enrollment while 
those without medical schools declined two per cent. The latter change 
occurred entin ly in the public sector, as Table 22A shows. 



92, 


,6 


98, 


,1 


100 


,0 


101, 


.7 


94, 


.4 


100, 


,2 


100, 


.0 


107, 


.9 



-59- 



Table 22 . Trends in Mean Enrollment for Advanced Degrees in Biological 
Sciences by Type of Institution 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



154 



163 



167 



172 



174 



175 



178 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All With Medical Schools 
All without Medical Schools 



49 


117 


123 


120 


122 


118 


119 


125 


28 


158 


166 


162 


165 


161 


160 


169 


21 


63 


66 


64 


65 


60 


64 


68 


71 


179 


190 


199 


206 


212 


215 


215 


30 


226 


238 


246 


254 


266 


271 


280 


41 


144 


155 


165 


172 


173 


174 


168 


58 


193 


203 


206 


211 


215 


217 


226 


62 


117 


125 


131 


136 


135 


136 


134 



Source: National Center for Education Statistics. 



Table 22A. Trends in Index Numbers for the Means of Enrollment for 

Advanced Degrees in Biological Sciences Shown in Table 22 
(Means for FY 1972 = 100) 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



87.6 92.6 95.2 98.1 99.0 100.0 101.7 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



49 


99.0 


103.5 


101.3 


103.3 


99.3 


100.0 


105.8 


28 


99.4 


103.8 


101.7 


103.6 


100.8 


100.0 


105.8 


21 


97.9 


102.8 


99.9 


102.1 


94.5 


100.0 


105.7 


71 


83.3 


88.5 


92.9 


96.2 


98.9 


100.0 


100.1 


30 


83.3 


87.7 


90.9 


93.6 


98.1 


100.0 


103.1 


41 


83.2 


89.4 


95.1 


99.0 


99.8 


100.0 


96.7 


58 


89.0 


93.4 


94.7 


97.2 


99.0 


100.0 


104.1 


62 


85.5. 


91.5 


95.9 


99.5 


98.9 


100.0 


98.1 



-60- 



Differences by Carnegie Commission categories of institutions . The data 
for advanced-degree enrollment in biological sciences for the Carnegie 
Commission groupings are presented in Appendix TABLES C-4 and C-4A. The 
most marked characteristic of the three private categories is the incon- 
sistency among their trend patterns. Research Universities I show an 
"oscillating" trend of increases and decreases (ending in FY 1973 at 
about its FY 1970 level, but with a seven-year increase of about 8 per 
cent, Research Universities II show an overall decline of 4 per cent, 
Other Categories show a generally upward trend, with a total increase 
of 11 per cent. 

For the public institutions: Research Universities I increase from an 
index of 85.4 to 102.2 (20 per cent); Research Universities II had an 
increase of 16 per cent; Other Categories increased 28 per cent. 

Advanced-degree enrollment in the health professions . The trend data in 
Tables 23 and 23A on the following page show two rather striking characteris- 
tics, for all institutions combined and for all of the sub-groups: (a) very 
high growth rates from FY 1967 through FY 1973; (b) generally steady increases 
from year to year (with very few "oscillations" in trend). 

The overall magnitudes of the graduate-enrollment increases over the seven 

years are as follows : 

All institutions combined 136% 

All private institutions 117% 

All public institutions 146% 

Institutions with medical schools 143% 

Institutions without medical schools 107% 

A comparison of trends in enrollment in health, profession and in federal 

R&D funding for medical sciences is shown in the following index numbers for 

all institutions combined: 

1968 1970 1972 1973 

Advanced-degree enrollment 50.0 62.8 100.0 110.7 
Federal R&D expenditures 110.6 107.8 100.0 106.2 



-61- 



Table 23 . Trends in Mean Enrollment for Advanced Degrees in Health 
Professions by Type of Institution 3 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



44 



47 



53 



59 



65 



94 



104 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



49 


41 


43 


48 


50 


57 


80 


89 


28 


58 


60 


69 


73 


84 


118 


129 


21 


18 


20 


21 


20 


20 


29 


36 


71 


46 


50 


56 


65 


71 


103 


113 


30 


89 


97 


112 


131 


144 


205 


227 


41 


13 


14 


14 


15 


17 


27 


28 


58 


74 


79 


91 


103 


115 


163 


180 


62 


15 


16 


16 


17 


18 


28 


31 



Source: National Center for Education Statistics. 



Table 23A. Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Health Professions Shown in Table 23 (Means for FY 
1972 = 100) 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



46.6 50.0 56.5 62.8 69.8 100.0 110.7 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



49 


51.2 


53.5 


60.7 


62.9 


70.8 


100.0 


112.0 


28 


49.3 


50.8 


58.8 


61.8 


71.4 


100.0 


109.9 


21 


60.9 


67.9 


70.7 


69.0 


67.9 


100.0 


123.3 


71 


44.2 


48.1 


54.3 


62.8 


69.2 


100.0 


110.0 


30 


43.3 


47.2 


54.7 


63.9 


70.5 


100.0 


111.1 


41 


49.1 


53.4 


51.9 


56.3 


61.9 


100.0 


103.6 


58 


45.4 


48.5 


56.1 


63.2 


70.8 


100.0 


110.7 


62 


53.4 


58.7 


58.7 


60.9 


64.1 


100.0 


110.8 



-62- 



The data for advanced^degree enrollment in the health professions by 

Carnegie Commission categories are given in Appendix TABLES C-5 and C-5A. Most 

of the enrollments are concentrated in Research Universities I and II, for both 

private and public universities. Their respective growth trends were quite 

different, however, as the index numbers in Table C-5A indicate. In overall 

percentage terms (FY 1967 through FY 1973) the differences are as follows: 

Research Univ. I Research Univ. II 

Private universities 166% 70% 

Public universities 106% 245% 

Advanced-degree enrollment in psychology . The means and index numbers 

for graduate psychology enrollments are shown in Tables 24 and 24A on the 

following page, with breakdowns by type of control and medical-school status. 

The comparative magnitudes of the overall enrollment increases from FY 1967 

through FY 1974 are as follows: 

All institutions combined 32% 

All private institutions 21% 

All public institutions 37% 

Institutions with medical schools 17% 

Institutions without medical schools ......... .52% 

Comparisons of advanced-degree enrollments in psychology for institutions 
classified by Carnegie Commission categories are shown in Appendix TABLES C-6 
and C-6A. Inspection of the index-number trends indicates that Research Uni- 
versities I, private and public, had the lowest rates of growth over the seven- 
year period, while Other Categories had the highest. The following overall 

percentages confirm these impressions: 

Research Research Other 

Universities I Universities II Categories 

Private universities 6% 28% 41% 

Public universities 215 32 70 



-63- 



Table 24 . Trends in Mean Enrollment for Advanced Degrees in Psychology by 
Type of Institution 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



72 



78 



86 



88 



88 



89 



95 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



49 


63 


65 


66 


68 


70 


69 


76 


28 


80 


77 


75 


76 


81 


79 


86 


21 


40 


50 


53 


56 


57 


56 


61 


71 


79 


88 


100 


103 


100 


103 


108 


30 


103 


113 


126 


133 


127 


127 


128 


41 


61 


69 


81 


81 


80 


86 


93 


58 


92 


95 


101 


105 


105 


104 


108 


62 


54 


63 


71 


72 


72 


75 


82 



Source: National Center for Education Statistics. 



Table 24A . Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Psychology Shown in Table 24 (Means for FY 1972 = 100) 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Institutions 



120 



81.4 



88.1 96.4 99.1 98.8 100.0 106.2 



All Private Institutions 
With Medical Schools 
Without Medical Schools 

All Public Institutions 
With Medical Schools 
Without Medical Schools 

All with Medical Schools 
All without Medical Schools 



49 


91.5 


94.9 


95.4 


98.1 


102.2 


100.0 


109.9 


28 


101.9 


97.8 


95.2 


96.6 


102.1 


100.0 


109.5 


21 


72.0 


89.4 


95.8 


100.9 


102.3 


100.0 


110.6 


71 


76.7 


85.0 


96.9 


99.6 


97.3 


100.0 


104.5 


30 


81.5 


88.7 


99.4 


104.6 


100.5 


100.0 


100.6 


41 


71.4 


80.9 


94.2 


94.2 


93.7 


100.0 


108.8 


58 


89.0 


92.1 


97.9 


101.6 


101.1 


100.0 


103.8 


62 


71.6 


83.0 


94.6 


95.9 


95.9 


100.0 


109.3 



-64- 



VII, TRENDS IN EARNED DOCTORAL DEGREES 

Statistics were available for each of the 11 years from FY 1964 through 
FY 1974 for earned doctoral degrees in all fields, with breakdowns by academic 
disciplines- in accordance with the NCES-HEGIS classification the number of 
institutions in the present sample having complete records for each of these 
years, however, was reduced to 102. The present chapter is limited to doctoral 
degrees in all fields combined and to those in the biomedical-behavioral fields. 
(M.D.'s, D.D.S.'s and similar professional doctorates were not included.) 
Trends in Earned Doctoral Degrees in All Fields Combined 

The mean number of doctoral degrees granted per institution and the corres- 
ponding index numbers for these means are shown in Tables 25 and 25A on the 
following page. 

Doctoral degrees by types of institutions . The number of earned doctorates 

increased significantly over the 11-year period from FY 1964 through FY 1974 

for all classes of institutions in the sample, as indicated by the following 

summary of overall percentage increases based on the means in Table 25: 

All institutions 220% 

All private institutions 167 

All public institutions 265 

All with medical schools 214 

All without medical schools 221 

Private vs. public universities . The index numbers in Table 25A show 
generally similar trends for private and public universities. However, 
the rate of growth was lower for the private than for the public sector — 
as indicated by the overall percentage increases from FY 1964 through 
FY 1974. 

Presence vs. absence of medical schools . There appeared to be no signifi- 
cant differences in the growth trends in earned doctorates over the 11- 
year period when private and public institutions with or without medical 
schools, respectively, were combined. Within the private and public 



-65- 



CTl 



IO 
CTi 



CM 

in 
CTl 



CTi 



O 

r- 
CTi 



O "1 i-t 


CTi In «3- 


CM tO 


o lo to 


i-H CM I-H 


•«»■ oo 


CM CM rH 


CM CS1 (VI 


cm i-h 



NHlO 


VO CN! CM 


CM VO 


ONN 


CM tO CM 


LO OO 


CM CM i-H 


CM (SI CM 


CM rH 



OHK1 
OlON 
CM CM i-H 



I-H CM O 
1SI CM CM 
INI CM CM 



oo 


en in vo 


WOlO 


tO vO 


o 


en m cm 


HNO 


*t- in 


CM 


HNH 


CM (SI (SI 


CM i-H 



\o 


OlNM 


NOlN 


VO CTl 


cti 


00 t CM 


O O CTl 


CM NO 


■H 


rH CM i-l 


CM CM rH 


CM i-H 



CTl 
vO 
CTl 



tOlfltv 


CO"! 


to O 


NNO 


r~- oo c~- 


O LO 


HNH 


i-H iH i-H 


CM rH 



C3 
O 



. CO 
vO 
CTl 



>o 

CTl 



vO 
VO 
CTl 



LO 

vO 
CTl 



vO 
CTl 



lo 


ivintv 


to o oo 


*»■ CTl 


lo 


"1 O CTl 


LO vO ** 


00 CM 


iH 


rH CM 


i-H iH i-H 


i-H iH 



tO 00 CM 


eoos 


LO o 


"» 00 CTl 


to «*■ to 


vO CM 



CM 

<N1 



1^ CM tO 

(si vo oo 



i-l vO 

t o 



vO 


CM O O 


OOIH 


lo in 


In 


00 O vO 


cn o t-~ 


CO vO 



vO 

vO 



LO CM CM 

i-n en lo 



OHOI 
vO vO LO 



tN lo 



0) 

£ 
D 
Z 



CM 


iflino 


I"- to »*■ 


oo *» 


o 


«S- CM CM 


LO (SI to 


«* LO 



a 
o 

<-!-( -H 

o +-> 

3 
<D 4-1 
CVH 

X+J 
H in 

a 





in 


in 


o 




in iH 


i-H 


in O 




c o 


in o 


rH43 




o in o 


C m O 


o o 




•h ih j: 


OH£ 


O CO 




-POD 


•H O O 


43 




3 OC0 


4-> O CO 


O rH 




4->43 


343 


CO a) 




■H O i-H 


4-> O rH 


u 


m 


4J CO 03 


•H CO « 


iH -H 


a 


in o 


4-> U 


wo 


o 


CH'H 


I/1H'H 


o a> 


H 


i— i ca "O 


C 03 T3 


•HS 


4-> 


U V 


l-H O O 


13 


3 


4>-H S 


•hS 


<u +j 


4-1 


4J T3 


OT3 


S 3 


H 


ca o +j 


•h a> «j 


o 


+J 


>S 3 


rH S 3 


43 4= 


m 


■H O 


£ O 


4-1 4J 


c 


r.4345 


3 43 43 


• H -H 


H— 1 


CWP 


a, 4J +j 


* * 




•H -H 


• H -H 




1-1 


HS* 


rH^S 


i-H rH 



<< 



LO 
CM 



43 
c« 
E- 

C 
•H 

s 

o 

43 
co 

V) 

•3 



tt. 





in 






a> 




a> 




M 




M 




4> 




a 




rH 




a 




u 




o 




4-> 




u 




o 




Q 




*3 




Q) 




a 




h 




o3 




W 




CH 




O 




in 


. 


C 


w 


03 


u 


0) 


•H 


S 


44 




in 


o 


•H 


43 


4-> 


*-> 


(fl 




4-> 


M 


co 


o 




<4H/— , 


a 


o 


o 


in o 


•H 


M rH 


4J 


a> 


03 


43 it 


u 


e 


3 


3 CM 


•a 


Z r- 


w 


Ol 




« rH 


H 


4) 


O 


•Or" 


4H 


CB. 




HH 


Ih 


H 


a> 


C O 


+j 


•HlH 


c 




4> 


in in 


U 


tj c 




a o3 


rH 


<u 0) 


03 


r. S 


C 


Hv^ 


o 




•H 




4-1 


• 


03 


< 




Z 


CM 




o 


r-t 




o 


XI 




H 


03 




3 


H 




o 




CO 







C7V 



CTi 



rsi 
ts. 
CTi 



CTl 



O 



CTi 
vO 
CTl 



v© 

CTl 



o 

CTl 



1/1 

vO 

CTi 



u 

e 

3 

z 



c 
o 

««H -H 
O 4-1 

3 

4> 4-> 
P..H 

X4-> 
H m 

C 



o">o CM If) o 



en 

Cl 


OlsiO 
0)0 


CTi CM t-~ 

OlOOl 
1^ 


cti en 
en en 


00 


CO 00 o 


tO 00 iO 


CM CM 


CM 

O 


to to to 

o o o 


CM •*■ O 

O o o 


^ rH 
O O 



o 
o 

i-H 


o o o 
o o o 

rH rH rH 


o o o 
o o o 

rH rH rH 


o o 

o o 

rH i-H 


to 


IflNOl 


* i-l IO 


*t 00 


00 

en 


en oo cm 

en en o 


f- to to 

en o en 


O CO 

o en 



f-vOrH tOtOlO tOO 



CM 

en 



CM 

oo 



^ CM O 

en en o 



rH *r en 
en en oo 



to CM 

en en 



00 00 00 



is. vO O 



o rH en 

00 00 t- 



oo oo 



00 00 CTi 
INtsts 



OOOlt 



en cm tN 

vO Is. O 



LO O 

tN tN 



O ** 



O CMrHLO CMIOCM 00CO 

VO tNfNtN vOvOvO vOvO 



VO rH LO 00 LO O 



IN- 


IO CM In 


tO tO TT 


00 IN 


LO 


vO vO <0 


LO L0 L0 


LO L0 



CTl CM ^S- rH L0 



LO 

to 


rH 00 00 


rH rH CM 

to to to 


LO O 

to to 


^ 1 


to to In 


rH «* en 


en eo 


l~l 

to 


tN LO CM 

to to ^t 


tN IN VO 
CM CM CM 


r-l O 

to to 



CM 


LO LO O 


In tO *!■ 


00 «» 


o 


^ CM CM 


LO CM IO 


^ LO 



C 

o 



in 

in rH 

C o 
O in o 

•HrH,C 

•P o o 
3 O CO 

•HUH 
4-> CO 03 
in u 

C r-i -H 
rH 03 "O 
U 0> 
CJ-H S 
4» T3 

cd D «-> 
>S 3 

■H O 

r.45 43 

O- 4-> *•> 



in 


O 


rH 


m o 


m O 


•-* 43 


C m O 


O O 


O rH43 


O CO 


•H o o 


43 


4-> O CO 


U rH 


343 


CO CS 


*J UH 


O 


■H CO 03 


rH -H 


W U 


03 T3 


tO rH -rt 


o o 


C 03-3 


•HS 


I-H U V 


T3 


•HS 


0) 4-1 


y-3 


S 3 


• H V 4-1 


o 


rHS 3 


43 43 


43 O 


4-> 4J 


3 43 43 


•H -H 


CL, 4J 4-1 


S * 



rH H^S HSJ 



< < 



-66- 



howeyar, there seemed to be conflicting trends which had the 
effect of cancelling each other so as to produce the apparent result of 
no difference for the contained figures. For example, private institutions 
without medical schools grew somewhat more rapidly after FY 1969 than those 
with medical schools; whereas the opposite was true for public institutions, 
But the magnitude of these variations was not great. 

Carnegie Commission categories of institutions . The trends in mean number 
of earned doctorates for institutions classified by three Carnegie Commis- 
sion categories are shown in Appendix TABLES D-l and D^IA. As would be 
expected, since the number of doctorates granted was a criterian used in 
creating this hierarchy, Research Universities I (both private and public) 
had by far the largest mean number of doctorates among the three groups 
used in the present study. 

In terms of growth rates, however, the percentage changes from FY 1964 to 
FY 1974 showed a markedly inverse relationship to the order of the groups 
in the classification hierarchy — as indicated by the following percentage 
increases between FY 1964 and FY 1974 in mean number of doctoral degrees 
granted: 

Private Public 

Institutions Institutions 
Research Universities I " 142% 203% 

Research Universities II 241 342 

Other Categories 295 615 

Comparison of doctorates with R&D revenues . From FY 1969 through FY 1974, 

earned doctoral degrees for all institutions combined increased, as already 

noted, while sponsored R&D revenues declined. The trends over the years common 

to the respective data series are shown in the following index numbers: 

1969 1970 1971 1972 1973 1974 

Earned doctoral degrees 82.8 92.7 98.3 100.0 102.8 99.6 

All sponsored R&D revenues 102.1 101.6 96.8 100.0 99.9 95.9 

Federally sponsored R&D revenues 106.2 104.2 96.4 100.0 98.7 94.5 

In terms of percentage change over the six-year period, the following are 

the comparative figures: earned doctorates, 20% increase; all sponsored R&D 

revenues, 6% decrease; federally sponsored R&D revenues, 10% decrease. 



-67- 



Trends in Earned Doctoral Degrees in BiomedjcalTBehavioyal Sciences 

It will be recalled that the composite category "biomedical-behavioral 
sciences" based on the NCES-HEGIS enrollment survey included biological sciences, 
health professions, and psychology. The same set of biomedical-behavioral dis- 
ciplines was included in the survey of earned doctorates and has been used in 
this analysis. 

Trend data will be reported first for all biomedical-behavioral fields 
combined. Then separate results will be presented for biological sciences and 
psychology. The numbers of doctorates in the health professions were so small 
that they will not be discussed, although tables showing means and index num- 
bers have been included in the text and in Appendix D. 

Doctorates in all biomedical-behavioral fields combined . The statistics 
on earned doctoral degrees are shown in Tables 26 and 26A on the following 
page. It should be noted that because of the small average number of doctorates 
granted per year in this category, rounding to whole numbers in Table 26 has 
sometimes produced apparent inconsistencies with the corresponding index numbers 
in Table 26A. The discrepancies may also have been due partly to the fact that 
the index numbers were computed as the means of the index numbers for individual 
institutions, rather than directly from the group means in Table 26. 

The number of doctorates granted in biomedical-behavioral sciences has in- 
creased at a substantially lower rate than the number of doctorates for all 
fields. The 11-year percentage changes for the classification of institutions 
by type of control and medical-school status were as follows: 



-68- 



cd 

o 

•H 
4> 

s 
o 



in 

4) 
41 
U 
00 
4Kd 
O C 

o 

Cd +-> 

U 3 
O -t-> 
•M -H 
u <-> 
O tf> 
Q C 

►H 

4> <+H 

c o 
r. 

CO a) 

PJ (X 

<4H E-. 

o 

S- X) 
0) 

m 

0) 

u 

a 



X> 

E 
3 
Z 



c • 
0) u 
4) CO 

I— t 

•H (h 

o 



■•a 



o> x: 



ca 
H 



Hh 






tO 

Cn 









O 
cn 



vO 



vO 
vo 

en 



u-1 

vO 



vD 
0> 



3 
Z 



C 

o 

Mh -h 

O 4-> 

3 

CD -M 
CVH 

x<-> 

t-> <n 

S 



C*~ rH *3- IO cm ot^ t-~ Cn 

to to *3- i-h ><f into t* cm 



r^ on w to Oicn vo o 
tO to •"* i-H «*■ ^-to «* to 



t~- 0> rH «*■ tO ChOl to O 

tO CM ^ f-4 «3- «»tO »»■ to 



r-~ cn cm ■*!■ noooi tno 
to cm *a- rH f «*to ** to 



Tt r-- oo to o to<o cm r-~ 

IO CMtOi-t ■* T*IO t N 



cm uivON oo i-Hin oo r- 

IO CMIOtH tOTttO tOCM 



00 Vrf H r-t VQ00 to CM 

CM CM tO rH tOIOCM tOCM 



LO CM CM o oocnvo OO 

CM CM tO rH CMCMCM IO CM 



CM 00 vO 00 tOf^tO r-~ 00 

CM HN CMCMCM CM rH 



O O-TJ-. 00 CM tn rH ^(O 

CM I-HCSI CMCMCM CMi-l 



oo r- to r-« otOr-~ ■>* to 

r-l iH CM CMCMiH CM i-i 



CM 


muio 


Mfl* 


oo •«■ 


O 


■«* CM CM 


in cm to 


«■ to 



C 

o 






I/) 
en •-< 

c o 

o «o 

•H i-l J2 

woo 

3 O CO 
4->X! 

•H O iH 

■u co at 
l/l o 

G -H-H 

w WO 

O 4> 

O-H S 

+J13 

id up 
>2 3 

■H O 

r.X! X! 

a. 4-> v 



01 

1-1 
in O 
C in O 

O rH J2 
•H O O 

HOW 
3 X! 

+J O iH 

•H CO ecj 
+-> O 

m iH .h 
C cd Ho- 
rn o 4> 

•hS 

O T3 
■H 4> 4-> 

-i2 3 
XI O 

3x:x; 

IX +-> -M 



V) 

iH 

o 

m o 

H£ 

o o 

O CO 

x; 

O iH 
CO Cd 

o 

fH -H 

edT3 

O 4> 
•HS 

2 3 

o 
xx: 



* * 



iH iHjES -HSS rHiH 

iH iH rH rH rH 

< < < << 



in 
u 



CO 

+j 

CO 



c 
o 



cd 
o 

3 
13 

UJ 

!-. 
O 

<+H 

r. 

4> 

*-> 

C 
4) 



cd 

C 

o 



cd 

z 



o 

(-. 

3 
O 
CO 
CO 



in 
a> 
o 

c 

D 
•H 

o 



cd 

r< 
O 

•H 
> 

cd 

X! 
4) 
03 



UJ 
O 

•H 

•o 

4) 

E 
o 

•H 

pa 



m 




4> 




4> 




^ 




CO 




« 




a 




rH 




cd 




r. 




O 




■M 




U 




o 




a 




T3 




<u 




c 




(-. 




cd 


r^ 


w o 




o 


<4H rH 


o 






u 


in 




C 


CM 


cd f» 


4) CTi 


z 


fH 


o>. 


Xttt. 


«J 






f- 


r< 


o 


0<4H 


MH 






<n 


in 


a 


r. 


» 


4> 


0) 


-OS 


e 


1 — 1 


3 




2 vO 




r-J 


X 




41 


0) 


13 rH 


ex> 


r-t 


re 




H 


C 




■H 


c 




•H 


<n 




T3 


c 


C 


S 


4) 


o 


r.X! 


E- CO 



c-~ 






CM 



o 
en 



en 
en 



00 

en 



vO 

cn 



cn 



cn 



cn 



fH 

4> 

X) 

s 

3 
Z 



c 
o 

<4H.H 
O 4-> 

3 

4) 4-> 
P.-H 
X4J 

t- m 
C 



o 


CM IO P» 


to * o 


l-» ^ 


CM 

o 

rH 


f» t^ vO 

o o o 

rH rH rH 


CT> CM \0 

cn o cn 

rH 


«■ 00 

o a< 


o 


CM ^ 00 


Nino 


*t * 


CM 

o 


to »j r-. 

o o o 


o o o 
o o o 


CM r^ 

o o 



o 
o 



o o o 
o o o 



o o o 
o o o 



oo 



fv fv r~ 

oo oo oo 



00 <» rH 

oo oo cn 



CM to 00 

CO00N 



CM tO CM 

r~ r~ r-~ 



oo 



vO oo oo 
t^ t— \0 



»S- cn oo 
\o m o 



m 
C 
O 

-H 



to 



in 
in r-l 

a o 

o m O 

•HtHXl 
4J O O 
3 O to 

+->x; 

>H UH 

■»-> co cd 
in u 

BH-H 
rH Cd TJ 

O 4) 
4>-H S 
4->T3 
cd 4) 4-> 

>S 3 
•H O 

r<x! x; 

P. 4-> 4-> 

•H .H 

rH* * 



in 

r-l 

m o 

C w 

oh x: 

•H O O 
+-> O CO 

3X! 

P UH 
•H CO Cd 
4-> O 

in rH -h 
C cd -3 

►H O 4) 
•H 2 
O T3 
•H 4> +J 
rHS 3 

X) o 

3X! X! 

0. +J +J 
•H -H 



o o 
o o 



o 
o 

rH 


rH CM O) 

o o cn 

|H rH 


O 00 CM 

o cn o 

rH r-i 


O rH 

o o 

r-t r-< 


00 


cn to m 


oo oo oo 


o m 


CM 

cn 


CM tO rH 

cn cn cn 


CM CM CM 

cn cn cn 


IO CM 

cn cn 



IO o 

oo cn 



oo to 



to \C 



00 00 
VO vO 



O 00 



cn 

LO 


to «* cn 
vO vO m 


oo vo cn 
to to to 


o cn 
o to 


LO 


00 \D t^ 


cm cn CM 


vO t 


to 


oo oo cn 

UO LO LO 


CM o to 

to to to 


to to 


rH 


cn to to 


o IO 00 


to VO 


O 
U1 


NOOl 
tO vO Tf 


VON* 
»* *»■ "» 


to to 

to ^- 



CM 


to to o 


f- to ^> 


oo ^t 


o 


«* CM CM 


tO CM tO 


«j- to 



in 

r-t 

o 
m O 

r-t X! 

o o 

O CO 

X! 

O rH 

CO cd 

o 

rH -H 

cd T3 
U 4) 

•H S 

-3 
4> +j 

2 3 
O 

XI Xi 

+-> +■> 

•H -H 



< < 



-69- 



All 102 institutions 106% 

All private institutions — ,.,.,.,...,.. 82 
All public institutions «....,., 110 

Institutions with medical schools 96 

Institutions without medical schools. .. .123 

Private vs. public universities . The trend patterns of index numbers for 
the several categories of institutions shown in Table 26A were generally 
similar through FY 1972. Thereafter, however, private institutions showed 
moderate increases in biomedical-behavioral doctorates through FY 1974, 
whereas there was little change in the public sector except for a decline 
for public universities without medical schools. 

Carnegie Commission categories of institutions . The trends in mean number 
of earned doctorates in biomedical-behavioral sciences for institutions 
classified by three Carnegie Commission categories are shown in Appendix 
TABLES D-2 and D-2A. 

The most striking result shown in these tables is the fact that public 
Research Universities I showed a substantial decline in doctorates, from 
an index number of 100 in FY 1972 to 93.1 in FY 1974. Since both of the 
other public categories showed increased during this period, the differ- 
ence between private and public institutions noted in the preceding sec- 
tion was due entirely to the decline in doctorates for public Research 
Universities I. 

Over the entire 11-year period, however doctorate production by the 
Carnegie Commission categories of public institutions increased at a 
greater overall rate than those for the corresponding private groups, 
as the following percentage changes from FY 1964 to FY 1974 show: 

Private Public 

Research Universities I 71% 75% 
Research Universities II 83 164 
Other Categories 225 300 

Comparison of trends in doctoral degrees with R&D expenditures in biomedi - 
cal-behavioral sciences . Whereas the trends for the number of earned doctorates 
and federally sponsored R&D revenues for all fields combined went in opposite 
directions, the corresponding comparisons for all biomedical-behavioral fields 
show generally similar patterns of increase for doctorate production and R&D 
expenditures. The following index numbers summarize the results for the years 
common to the three variables listed: 



-70- 



1964 1968 1970 1972 1974 

Earned doctoral degrees 50.1 75.9 92.8 100.0 102.8 

Total sponsored research expenditures 75.7 95.8 100.4 100.0 109.0 
Federally sponsored research expenditures 76.9 103.3 105.6 100.0 107.6 

It is particularly interesting that both total and federal R&D expenditures 
increased at considerably more rapid rates between FY 1972 and 1974 than did 
doctorate production. (Such direct comparisons should be made in recognition 
of the fact that a time lag would be expected to intervene between funding in- 
put and doctoral output if a causal relationship between these two variables 
is assumed to exist.) 

Doctorates in biological and psychological sciences . ° It will be conven- 
ient to discuss the data on earned doctorates in the biological sciences and 
in psychology together, partly to emphasize the marked contrast in trends be- 
tween the two fields from FY 1972 to 1974. But first the overall percentage 
changes in doctorates granted from FY 1964 through FY 1974 will be compared by 
types of institutions. (The percentages shown earlier for all biomedical- 

behavioral fields will also be reproduced.) 

Biomedical-Behavioral Biological Psychology 
Sciences Sciences 

All 102 institutions 106% 
All private institutions 82 

All public institutions 110 
Institituions with medical schools 96 

Institutions without medical schools 123 

Trend data are shown in Tables 27-27A (biological sciences) and Tables 29- 

29A (psychology) . The two sets of index numbers (Tables 27A and 29A) differ 

in one major respect between FY 1964 and FY 1972: the biological sciences 

began at a lower level in FY 1964 (relative to their FY 1972 base) and hence 



100% 


117% 


89 


71 


117 


133 


86 


78 


100 


150 



It was noted above that too few doctoral degrees were granted in the health 
professions to justify any discussion of trends in these fields. But trend data 
are shown for them in Tables 28 and 28A and in Appendix TABLES D-4 and D-4A. 



-71- 



co 










r-» 


co 


r» -a- oo 


oo o nO 


r-» O 


ON 


CM 


rH CM 


CM CO CM 


CM CM 



u 
cu 

•i 



c 
o 

4H -H 
O 4J 
3 
CU 4-> 
0.-H 
>N 4J 

H (0 

C 



r^ in oo 


vO 00 -J- 


no oo 


rH CM 


CM CM CM 


CM rH 



vr 


oo in 00 


00 rH NO 


00 O 


CM 


rH CM 


CM CO CM 


CM CM 



00 NO 00 


COHvO 


OO o 


rH CM 


CM CO CM 


CM CM 



CM 


\CP1N 


no on in 


no oo 


CM 


i-l CM 


CM CM CM 


CM rH 



m o oo 


m r» -a- 


co oo 


rH CM 


CM CM CM 


CM rH 



m cn r~ 


rH CO o 


cm m 


i-H CM 


CM CM CM 


CM rH 



oo oo oo 

rH rH rH 



NO O 



CM 


m m o 


P*. CO -a 1 


oo -a 


O 


•a- CM CM 


m CM CO 


-a m 


rH 















co 






CO 







to 




rH 






rH 


CO o 




c 




O 


co 







rH Xi 




o 


to 


O 


c 


CO 





o u 




•H 


rH 


x: 


o 


rH 


rC 


o oo 




■u 





o 


•H 





o 


xt 




3 


o 


CO 


U 





CO 


CJ rH 




4-> 


X 




3 


x 




CO co 




•H 


U 


rH 


4-1 


o 


rH 


u 


in 


4-1 


cn 


a! 


•H 


cn 


CO 


rH -H 


c 


ID 




u 


4-1 




CJ 


CO TJ 


r> 


a 


H 


•H 


CO 


rH 


•H 


CJ H) 


H 


M 


CO 


tj 


c 


a 


13 


•H g 


4-1 




c; 


01 


M 


t; 


CU 


TJ 


3 


0) 


•H 


g 




•H 


R 


0) 4-j 


u 


4-1 


T1 




o 


TJ 




g 3 


•H 


CO 


HI 


4J 


•H 


01 


4-1 





u 


> 


T, 


3 


rH 


•r. 


3 


J= rC 


C/l 


■H 




o 


X 




O 


4J 4-1 


e 


M 


■C 


rC 


3 


x 


■C 


•H iH 


-H 


Ph 


4-1 


4-> 


P- 


u 


4-1 


3 3 






■H 


•H 




•H 


•H 




H 


rH 


s 


3 


rH 


rs 


3 


rH rH 


rH 


3 






rH 
< 






33 



4J 



(0 
o 

3 
TJ 
W 

t-i 

o 



u 

CU 
4-> 

c 

CU 

o 



CO 



01 

o 
l-l 

3 

o 
cn 



o 

X 

oo 

CO 
CU 

o 
c 

CU 

•w 
u 
oo 



CD 
CJ 
•H 
00 

o 

rH 

o 

iH 

oa 

c 



CO 
CU 
CU 

1-4 

00 
cu 
Q 



M 
o 

4J 

cj 
o 
o 

tj 

cu 



CO 

w 



CO 


o 


c 


o 


CO 


rH 


01 




S 


II 


<u 


CM 


rC 


r-» 


«J 


ON 




rH 


l-l 







£ 


cn 


S-4 


i-i 





cu 


IIH 


-a 




1=1 


CO 


3 


c 


!5 


CO 




01 


r. 


>~ 


(U 




XI 




c 





•H CU 
rH 

CO Xl 

TJ CO 

C H 
CU 

l-i e 

H -H 



S-i 
CU 

■§ 

3 
5S 



c 
o 

UH t-l 
O 4J 

3 

CU 4J 

P.-H 

S>n 4-1 

H to 

C 



CO 

On 



00 
On 



O 

o 



o 
o 



no r-~ cn 

ON ON On 



rH CM rH 
ON ON ON 



NO VO NO 
ON ON ON 



o o o 



oo rso 

ON ON O 



o o o 
o o o 



CM CM rH 

o o o 



rH r~ r^ 



o o o 
o o o 



O ON rH 

O On O 



00 O <N 



no in on 
in in m 



oo no oo 
m in in 



rH 3 3 



rH 3 3 



<t rH 
ON ON 



r~ on 

ON ON 



o o 
o o 



O rH 

o o 



CM 

ON 


rH 
ON 


rH 00 
ON 00 


CO 
ON 


ON 


CO 
ON 


CM 

ON 


CO 

ON 


CN 


ON 


O On 


00 


r^ 


o 


NO 


o 


p~ 

co 


rH 
00 


ON CN 
P*- ON 


ON 

00 


m 

00 


CO 
ON 


CM 

oo 


CO 

ON 



oo 


m 

00 


oo 


CO 
p^ 


-a 
p~ 


co 


NO 


ON 


NO 


m 


m 


CM 


CO 


o 


r- 


CM 


rH 


co 


NO 


nD 


ON 

NO 


rH 
NO 


CO 

NO 




P^ 

NO 


CO 

NO 


NO 

nO 



no 00 

m m 



•3- 

m 


in 

in 




NO 

in 


m 


NO 

in 


CM 

m 


in 
m 




o 


p^ 


r— 


CM 


rH 


r^ 


<r 


00 


NO 


p-- 
-3- 


CM 

in 




m 
-d- 




ro 


•a- 


00 

-a- 





CM 


m m o 


r^ co ^a 


oo >a 


O 


-a CN CM 


in cm co 


<a m 









CO 






CO 


o 




CO 




rH 






rH 


to o 




a 




O 


CO 







rH X 







CO 


a 


c 


CO 


O 


O CJ 




•H 


rH 


x; 


o 


rH 


X 


O C/l 




4-1 





CJ 


•rl 





CJ 


X 




3 


a 


00 


4-1 


o 


C/l 


CJ rH 




4J 


rC 




3 


X 




00 co 




•H 


t,J 


rH 


4-1 


o 


rH 


CJ 


CO 


*J 


C/l 


CO 


•H 


m 


CO 


rH -H 


C 


CO 




CJ 


4-1 




CJ 


CO TJ 





B 


rH 


■H 


co 


rH 


•H 


CJ CU 


■H 


M 


cn 


TJ 


S 


CO 


TJ 


•H g 


4-1 




CJ 


CU 


M 


CJ 


CU 


TJ 


3 


CU 


•H 


g 




-H 


g 


CU 4-> 


4-1 


4-1 


■n 




U 


X) 




g 3 


H 


CO 


01 


4-1 


•rl 


•U 


4-1 


O 


4-1 


> 


*: 


3 


rH 


:>-: 


3 


x: xi 


IT 


•H 




o 


Xi 







4-1 4J 


C 


l-l 


XI 


X 


3 


X 


X 


i-I -H 


M 


Ph 


4J 


4-1 


P-i 


a 


4-1 


3 3 



< < 



-72- 



co 
er. 



HMO 



00 
NO 

as 



r- 
vO 

ctn 



NO 
OS 



m 

NO 

ctn 



•9 
vO 
On 



HHO 



i-H i-H O 



rH iH r-l O 



co m cm -a- iH 



co «r cm co 



nuiN n h 



CO in CM ■* H 



CM CO CM CO rH 



CM CO CM iM H 



CM CM rH CM i-l 



CM CM CM CM rH 



CM CM CM CM rH 



CM CM rH rH i-H 



CM CM rH rH rH 



H 

| 

•z. 



CM 


m in o 


t-* co »3- 


oo -a- 


O 


-» CM CM 


m cm co 


-a- m 



e 
o 

O 4J 

&£ 

H m 
(3 





01 


rH 




B 


O 




O 


CO o 




iH 


rH J= 




■U 


O u 




3 


O lO 




4J 


J= 




•H 


O rH 


(0 


4J 


00 CD 


C 


(0 


O 


o 


c 


rH TH 


iH 
4J 


M 


(9 -O 

o 0) 


3 


01 


•H £ 


4-1 


4J 


X> 


•H 


CO 


01 4J 


•U 


> 


£ 3 


CO 


iH 


o 


c 


M 


j= x: 


M 


P* 


4-> 4-1 



rH 3 3 
3 



0) O 

e <o o 

O rH 4= 

iH O O 

4J O CO 

3 x: 

4J O rH 

•rl CO CO 

4J CJ 

0) rH TH 

G CO T3 

M O 01 

•H £ 
O TJ 

iH 01 4J 

rH 2 3 

XI o 

3 x: x: 

Cu 4J 4-1 

•H -H 

■H 3 3 

3 



to 

rH 

o 
to o 

rH J3 

o o 

O CO 
X! 

o <-t 

co co 

o 

rH tH 

CO TJ 

0) 
•H £ 

-o 

01 *-> 

x § 

j= x: 

.'J 4-t 
■H -H 

3 3 



< < 



10 

o 



CO 
u 

CO 



CO 
O 

3 
-O 
W 

o 



M 

o> 

4J 

c 

01 

o 



CO 

5S 



0> 

o 

)H 

3 
O 
CO 



o 
x: 

CO 

to 
c 
o 

•H 
CO 

0) 
01 

MH 

o 
u 



CO 
01 

32 



(0 
01 
01 

u 

t>0 
01 

Q 



CO 
>-< 

o 

4-1 
CJ 

o 

Q 






38 

CO rH 
01 

£ n 

01 CM 

jr, r» 

4-1 CTN 



(-1 

o 



10 S4 

rl O 
01 <4H 
X> 

to 

3 a 

55 co 
0) 

x £ 

T3 

c 

H 00 
CN 
C 

•iH 0) 
rH 
10 J3 
•O CO 
C H 
01 

£5 



r~ 


• 


• • 


• • ■ 


. . 


on 


oo 


<>o<o 


oi h in 


CO 1^" 


r-l 


o 


CO CO rH 


o>oo> 


•-I on 



u 

01 

•§ 

3 

z 



rH 


r** ctn co 
-3" -a CO 

rH rH rH 


r- no 

o o 

r-t rH 


CTn 
O 

rH 


o 

CM 
.-H 


rH 
rH 


CM 


CO GO CO 


*d" co 


m 


ON 


rH 


CO 
ON 


rH rH CO 

O OM» 


O CO 
ON On 


m 
oo 


CM 

ON 


01 



o 
o 



nO 



o o o 
o o o 



o o o 
o o o 



rH 00 CTn 

CTn H \D I 

-3- <r <f 



CO \D 
CM CM 



rH no r-^ 
in co r-~ 



^-i r-\ r» 

in sj vO 



O 

o 



nO 

CO 


CO 


m 

ON 


CO 
CO 


m 

00 


r-i 


oo 
o 

rH 


CTN <-t 

t~~ o 


NO 


m 


m 




CO 


o 


CTN 


no m 


NO 


m 

NO 


CO 


1 


00 


ON 

m 


CO 

oo 


CO NO 

no r^ 


CO 


CTn 


rH 




r^ 


CO 


r-t 


CM CO 


no 

m 


O 

m 


in 


1 


CO 

m 




NO 
NO 


in O 
m vO 



CTN 


rH CTN 1 


CM -3- CO 


CTN -3- 


NO 


NO NO 


r~ in o 


m ctn 



in 


rH NO 1 


CO CM CTN 


o o 


NO 


-tf -* 


r~ m o 


in o 



CTN O 
CO P^ 



r- ^-t 
CO NO 



m m o 


r~- co -a 


oo -3- 


-a CM CM 


m cm co 


<r m 











01 






0] 


o 






CO 




r-t 






rH 


01 O 






C 







to 




O 


<-i jr, 









01 


o 


c 


01 





o o 






•H 


rH 


-C 


o 


rH 


JC 


o co 






4J 


o 





•rl 


O 


CJ 


J= 






3 





CO 


4-1 





CO 


U rH 


a 




U 


£ 




3 


-C 




CO CO 


o 




■rl 


u 


rH 


4-1 


CJ 


rH 


o 


UH -H 


CO 


4-1 


to 


CO 


•H 


CO 


CO 


<-l iH 


O 4J 


a 


to 




CJ 


4-1 




o 


CO tl 


3 


o 


c 


rH 


•H 


01 


rH 


•rl 


O 0) 


01 4-1 


•H 


rH 


CO 


XI 


C 


CO 


TJ 


■H £ 


a. -h 


U 




CJ 


01 


M 


o 


01 


•a 


>< u 


3 


01 


■H 


S3 




iH 


£ 


01 4-1 


H to 


4J 


4-1 


TD 




U 


-D 




£ 3 


c 


iH 


CO 


0) 


4-1 


•H 


0) 


4J 


o 


M 


4J 


> 


£ 


3 


rH 


S 


3 


jr. xi 




CO 


•H 




O 


XI 







4J 4J 




C 


M 


J= 


J= 


3 


X 


X 


•H TH 




rH 


Ol 


•H 


4-1 
•H 


Ph 


*H 
tH 


4J 

•H 


S 3 




rH 


rH 


3 


3 


i-H 


3 


3 


rH •-< 




rH 


rH 






rH 






<-t rl 




< 


< 






<C 






< < 



-73- 



co 
a 

CO 



oo 
so 



SO 
OS 



o 

sO 
Os 



ul 
sO 

o\ 



sO 

Os 



U 
0) 

j 

3 



c 
o 

4-1 

3 

01 4-> 

>» x> 

H 01 

a 



ei cm sc r» -*r~rH voo 

rH i-li-( rlHH rH rH 



CM N*N CM m rH sO Os 



o -a- m n n h oo o* 



■H o co m n n h no\ 



O onm rH co cy. oo co 



o O-3-u-l onoi co oo 



00 00 rH U0 CO H I-> rH SO 



oo as co -a - eooir- h vo 



r-» oo i— I oo vor^vo as in 



so r-. on *3* so so in oo m 



SO S O fl sO 00 *3° ON -3* 



in in o 


r-~ co -r 


00 sj- 


<• CM CM 


in cm co 


•* in 









to 






CO 


o 




CO 




i-H 






l-l 


co o 




C 







CD 




o 


■H 43 




o 


CO 


O 


C 


0) 


o 


o o 




•w 


rH 


43 


o 


r-l 


43 


O CO 




4J 


O 


u 


■H 


O 


u 


43 




3 





CO 


•u 





CO 


CJ rH 




•u 


43 




3 


4= 




CO CO 




■H 


U 


l-l 


4J 


CJ 


rH 


u 


CO 


4-1 


on 


CO 


■H 


C/l 


CO 


rH -H 


a 


CO 




o 


4J 




CJ 


CO "O 


o 


c 


H 


•H 


CO 


rH 


•H 


o «J 


•H 


M 


a 


-O 


3 


cti 


*o 


•H g 


■U 




u 


01 


M 


a 


a> 


T) 


3 


0) 


■H 


S 




-H 


?: 


01 U 


■u 


4J 


T3 




u 


•a 




X 3 


■H 


CO 


ID 


4-> 


■H 


<u 


■u 


O 


U 


> 


H 


3 


r-l 


B 


3 


X X 


CO 


•H 




O 


* 




o 


4-1 4-1 


c 


U 


43 


4= 


3 


43 


43 


•H -H 


H 


ex 


4-> 


4J 


IX 


4J 


4-1 


3 3 



^ 3 3 



rH 3 3 rH rH 



CO 
4J 
CO 



CO 

o 

3 
T3 

W 

S-> 

o 



cu 
4-1 

c 
a> 

CJ 



CO 

z 



cu 
o 
l-l 

3 

o 

CO 



o 

43 

CO 

>> 

00 

o 

rH 

o 
43 
o 
>> 

CO 

IX 



CO 

CU 

(11 

1-4 

00 
01 

o 



CO 
l-l 
o 

4-1 

o 
o 

Q 

•o 

01 

c 

u 

CO 

w 



to 




C 




rt 


*■** 


CU 


o 


S c 




rH 


<u 




45 


II 


4-J 






CM 


M 


r— 





on 


U-i 


rH 


CO 
!-4 


£ 


cu 




42 


u 


H 


o 


3 


uh 


Z 






CO 


« 


C 


01 


CO 



C X 



CO 

TJ CU 

C rH 

4) jo 

I-" CO 

H H 



l-l 

cu 
1 

3 
Z 



c 
o 

HH -H 
O 4-> 

3 
CU 4-) 
CX -H 
>N 4-1 

H co 

c 



rH OS sO 
CM rH CM 



m in r-~ 

rH CM O 



ON 
ON 



\0 in in 

OS Os OS 



rH so in 
o as o 



so m as 

ON Os OS 



co as rv 
os OS oo 



-3- 

SO 



oo 
in 



r> CM CM 
f-~ 00 SO 



-3- m <■ 
m m in 



rH Cv rH 00 



sO sO CO 
so so sO 



C- -3 OS 

<f <f -I- 



rH m oo 
n rv in 



Os rv i— t 
-* m <- 



rH O 
CM CM 



rH 
rH 

rH 


o 

CM 

rH 


o 

CM 


rH 
CM 
rH 


m 
o 

.-i 


o 

rH 


rH 

o 

rH 


m 

rH 

rH 


SO 

o 

rH 


o 


O 


O 


o 


o 


o 


o 


o 


o 


o 
o 


O 

o 


O 

o 


o 
o 


o 

o 


o 
o 


o 

o 


o 
o 


o 
o 



r- co 
os o 



r- O 

Os OS 



CO 
OS 


Os 
Os 


o 

rH 


oo 


00 

oo 


00 


OS 
oo 


as 


CO 

oo 


in 


oo 


sO 


-* 


CO 


in 


<!■ 


sO 


in 




o 

00 


as 


-3- 
00 


o 


as 
rv 


CM 

sO 


cr> 


sD 


SO 


OS 


CM 


sO 


rs 


o> 


CO 


CO 


rH 


sO 

rv 


CM 
C^ 


sO 

as 


CM 
00 


in 
so 


sO 


SO 

SO 


rH 
00 


o 



OS so 

so in 



m <r 
m m 



so m 
so -3- 



m m o 


Pv CO -3 


CO <T 


-3 CM CM 


in CM CO 


<r m 









CO 






CO 


o 




CO 




rH 






rH 


co o 




c 




o 


CO 




o 


rH 43 







ID 


o 


C 


CO 


O 


O O 




iH 


rH 


43 


o 


rH 


43 


O CO 




4-4 


o 


CJ 


•H 


CJ 


CJ 


43 




3 





CO 


4-1 





CO 


CJ rH 




4-1 


43 




3 


43 




CO CO 




•H 


CJ 


rH 


4-1 


CJ 


rH 


CJ 


CO 


4-1 


U1 


CO 


•H 


cn 


CO 


rH -H 


C 


CO 




U 


4J 




CJ 


CO T3 


o 


3 


H 


•H 


CO 


rH 


■H 


CJ 01 


H 


M 


cr 


TJ 


a 


cfl 


■a 


■H S 


4-1 




') 


CU 


M 


CJ 


01 


"O 


3 


01 


■H 


r. 




•H 


s 


01 4-1 


4J 


4-> 


-a 




U 


-a 




S 3 


H 


CO 


i) 


4-1 


•H 


V 


4-) 


O 


4-1 


> 


X 


3 


rH 


X 


3 


43 43 


CO 


•H 




O 


43 




O 


4-1 4-1 


C 


u 


43 


43 


3 


43 


43 


•H tH 


-1 


ex 


4J 


4-1 


Q-i 


4-J 


4-1 


3 3 






■H 


T-l 




•H 


•H 




rH 


rH 


3 


3 


rH 


■Ji 


3 


rH rH 



-74- 



had a somewhat more rapid growth rate than psychology. But within each disci- 
pline, the index numbers for the various sub-groups show fairly consistent 

increases, although numerous irregularities occurred in individual years in 
both tables. 

The principal differences between the two fields, however, is that the 
mean numbers of doctorates granted in psychology increased markedly after FY 
1972 while those in the biological sciences declined. For all institutions com- 
bined, the increase for psychology was 17.9 per cent whereas the biological 
sciences decreased 6.8 per cent during the two-year period. These divergent 
trends were rather uniformaly reflected in the indices for all of the sub-groups 
shown in Tables 27A (biological sciences) and 29A (psychology) . 

Comparisons of doctoral trend data for these two sets of disciplines, by 
the Carnegie Commission classification of institutions, are shown in Appendix 
TABLES D-3 and D-3A (biological sciences) and TABLES D-5 and D-5A (psychology). 
The overall increases between FY 1964 and 1974 are summarized in terms of the 
following percentages: 



Private Institutions 

Research Universities I 
Research Universities II 
Other Categories 

Public Universities 

Research Universities I 
Research Universities II 
Other Categories 



Biological 


Psychology 


Sciences 




75% 


42% 


28 


140 


300 


150 


80 


58 


110 


225 


267 


300 



-75- 



VIII, TRENDS IN THE FUNDING PATTERNS OF NIH AND ADAMHA 

In the recent report of the National Science Board to the President, 9 it 
is stated that biomedical research accounts for some 90 per cent of the federal 
obligations for health-related R&D in FY 1974, the greater part of which is 
funded by the nine National Institutes of Health. A second category, mental 
health, is reported to account for five per cent of the remainder — all of which 
is the responsibility of ADAMHA's National Institute of Mental Health. NIH 
and ADAMHA account for almost all federally funded "biomedical-behavioral" 
R&D, which in turn is the source of funds for most academic research in these 
fields. 

In addition to R&D support, these two agencies also provide funds to uni- 
versities for several other purposes; and changes in the distribution of their 
awards among funding categories can have serious effects upon institutional 
programs. The purpose of this chapter is to analyze trends in the funding 
patterns of NIH and ADAMHA and in the levels of support for the various cate- 
gories of awards over the seven-year period for which data were available. 

Computerized records of the funding obligations for both agencies since 
1969, have been maintained in NIH's IMPAC files, and data from these files have 
been secured for 145 of the 148 universities involved in the present study. 



" Science Indicators, 1974 , The National Science Board/National Science Founda- 
tion, 1975. 



-76- 



Three types, of trend analysis have been used; (a) percentage distributions by 
, type of funding mechanism for each of the seven fiscal years ; (b) trends in the 
mean amounts awarded per university for each category of award over the seven 
fiscal years; (c) trends in the respective percentages of funds awarded for 
direct and for indirect costs by funding mechanism. 
Trends in the Distribution of Awards by Funding Mechanism 

The trends in percentage distribution of awards by funding mechanism are 
shown for NIH in Table 30 and for ADAMHA in Table 31 on the following page. 

NIH funding distributions . The percentages in Table 30 show that NIH 
awards have gone predominantly for regular research grants to individual 
faculty members. The proportion of the total amount awarded in this category 
has increased steadily from 59 per cent in FY 1969 to 71.2 per cent in FY 1975. 
(The maximum was 71.9 per cent in FY 1973.) 

The category with the next largest proportion was training grants, which 
has shown a decline from 18.6 per cent of the total in FY 1969 to 11.9 per cent 
in FY 1975. The related category of fellowships has also followed a predomin- 
antly downward trend: from 8.3 per cent in FY 1969 to a low point of 2 per 
cent in FY 1973, followed by an increase to 3.8 per cent in FY 1975. 

There has been no appreciable change in the relative level of support for 
"program project grants," whose proportion of NIH awards to the 145 universities 
• increased from 8.1 per cent in FY 1969 to 11.8 per cent in FY 1972, followed 
by a decline to 8.7 per cent in FY 1975. 



-77- 



Table 30. Trends in Percentage Distributions for NIH Awards by 



Funding 


Mechanism 


to 145 i 


Jnivers: 


Lties a 








Funding 






Fiscal Year 






Mechanism 


1969 


1970 


1971 


1972 


1973 


1974 


1975 


Regular Research Grants 


59.0% 


59.7% 


61.9% 


65.5% 


71.9% 


70.3% 


71.2% 


Program Project Grants 


8.1 


9.6 


11.0 


11.8 


10.0 


8.5 


8.7 


Clinical Research Grants 


1.2 


1.2 


1.3 


0.4 


0.4 


0.3 


0.3 


Training Grants 


18.6 


18.0 


15.8 


14.9 


11.4 


14.2 


11.9 


Faculty Awards 


3.0 


3.3 


3.3 


3.1 


3.4 


2.3 


2.6 


Fellowships 


8.3 


6.3 


5.2 


3.1 


2.0 


3.2 


3.8 


Other Awards 


1.8 


1.9 


1.5 


1.2 


0.9 


1.2 


1.5 


All Categories 


100.0 


100.0 


100.0 


100.0 


100.0 


100.0 


100.0 



Table 31 . Trends in Percentage Distributions for ADAMHA Awards by 
Funding Mechanism to 145 Universities 3 ^ 



Funding 

Mechanism 



Fiscal Year 



1969 1970 1971 1972 1973 1974 1975 



Regular Research Grants 
Program Project Grants 
Clinical Research Grants 
Training Grants 
Faculty Awards 
Fellowships 
Other Awards 
All Categories 



35.6% 


33.3% 


36.1% 


38.1% 


44.3% 


39 . 8% 


42.0% 


3.2 


4.3 


4.3 


4.1 


3.9 


5.2 


2.4 


44.2 


43.8 


44.0 


42.5 


42.2 


45.5 


39.3 


2.5 


2.7 


2.5 


2.2 


2.1 


1.4 


1.4 


11.0 


11.5 


7.9 


8.0 


3.1 


3.0 


8.9 


3.5 


4.4 


5.1 


5.1 


4.4 


5.1 


6.0 



100.0 100.0 100.0 100.0 100.0 100.0 100.0 



Source: National Institutes of Health, IMP AC files. 

It should be noted that ADAMHA' s training grants are made largely for clinical 
training in such areas as psychiatry, clinical psychology, psychiatric nursing, 
and paraprofessional training. Unlike NIH training grants, only a small pro- 
portion of ADAMHA' s grants in this category are for research training. 



-78- 



ADAMHA funding distributions . It is eyident from the data in Table 31 
that ADAMHA has had quite a different pattern of distribution of awards to 
universities from that shown by NIH, Regular research grants have received 
a far smaller proportion of ADAMHA' s total awards to universities — the 
indices ranging from 35.6 per cent in FY 1969 to 42 per cent in FY 1975. 
The highest proportion of ADAMHA's university funding has gone for training 
grants (heavily for clinical training in mental health specialities) — the per- 
centages remaining fairly stable over the seven-year period (44.2 per cent in 
FY 1969 and 39.3 per cent in FY 1975). 

Relatively more of ADAMHA's funds have gone for fellowships than was the 
case with NIH, but the trends for the two agencies have been generally similar. 
ADAMHA's fellowship percentage declined from 11 per cent in FY 1969 to 3 per 
cent in FY 1973 but then jumped to 8.9 per cent in FY 1975. 
Trends in Amounts of NIH Awards to Universities by Funding Mechanisms 

The percentage distributions of funds by years give no indication of the 
levels of support in dollars provided to the institutions for the various types 
of funding mechanisms. In order to determine the trends in the amounts of funds 
awarded, averages (means) have been computed in constant dollars for each funding 
category over the seven-year period for all institutions combined. The results 
for NIH are shown in Table 32 on the following page. From the mean amounts in 
Table 32, index numbers have been computed for each funding category — using 
the mean for FY 1972 as the base — and these indices are shown in Table 32A. 
For example, the mean total award per institution for regular research grants 

in FY 1969 was $326 thousand in constant dollars, and by FY 1975 the mean had 



-79- 



Table 32 . Trends in Mean NIH Awards to 145 Universities by Funding 

Mechanism — Constant Dollars in Thousands (NIH R&D Deflator, 
FY 1964 = 100) a 



Funding Fiscal Year 

Mechanism 1969 1970 1971 1972 1973 1974 1975 



Regular Research Grants $325.6 $301.6 $331.0 $369.9 $361.2 $427.8 $439.8 

Program Project Grants 44.8 48.5 58.8 66.4 50.2 51.7 53.8 

Clinical Research Grants 

Training Grants 

Faculty Awards 

Fellowships 

Other Awards 

All Categories 551.4 505.4 534.0 564.8 502.4 608.5 617.3 



6.5 


6.2 


6.8 


2.0 


2.1 


2.1 


1.8 


02.3 


91.2 


84.5 


84.3 


57.5 


86.1 


73.5 


16.7 


16.8 


17.8 


17.5 


17.0 


14.0 


16.3 


46.0 


31.7 


27.8 


17.7 


9.8 


19.3 


23.2 


9.6 


9.4 


8.0 


6.9 


4.7 


7.5 


8.9 



a Source: National Institutes of Health, IMPAC files. 

Table 32A. Trends in Index Numbers for the Means of NIH Awards 
Table 32 (Means for FY 1972 = 100) 


Shown in 




Funding Fiscal Year 


Mechanism 1969 1970 1971 1972 1973 


1974 


1975 



Regular Research Grants 

Program Project Grants 

Clinical Research Grants 

Training Grants 

Faculty Awards 

Fellowships 259.2 178.6 156.9 100.0 54.9 108.8 130.5 

Other Awards 139.4 136.0 115.5 100.0 68.2 109.2 128.2 

All Categories 97.6 89.4 94.5 100.0 88.9 107.7 109.3 



88.0 


81.5 


89.3 


100.0 


97.6 


115.6 


118.8 


67.4 


73.1 


88.4 


100.0 


75.5 


77.7 


81.0 


325.3 


312.4 


341.7 


100.0 


103.0 


104.1 


90.9 


121.2 


108.0 


100.2 


100.0 


68.1 


102.1 


87.2 


95.2 


95.6 


99.7 


100.0 


97.1 


79.8 


93.1 



-80- 



increased to $440 thousand (a gain of 35 per cent in real dollars). In the case 
of training grants, on the other hand, there was a decrease from $102 thousand 
in FY 1969 to $73.5 thousand in 1975 (a decline of 28 per cent). 

For all categories combined, NIH had a relatively small increase in average 
amount per university from $551 thousand in FY 1969 to $617 thousand in FY 
1975 (12 per cent). 
Trends in Amounts of ADAMHA Awards to Universities by Funding Mechanisms 

Comparisons of ADAMHA funding mechanisms in terms of trends in the mean 
amounts of awards are shown in Tables 33 and 33A on the following page. By 
contrast with NIH funding levels, almost all of the ADAMHA funding categories 
showed declines over the seven-year period. For example, the mean amount 
awarded for regular research grants declined from $78.2 thousand in FY 1969 
to $68.8 thousand in FY 1975 (a decrease of 12 per cent). Training grants 
similarly declined from a mean of $96.9 thousand in FY 1969 to $64.4 thousand 
in FY 1975 (a drop of 34 per cent) . 

Unlike NIH, ADAMHA suffered an overall decline in constant dollars in its 
awards to this sample of universities: from $219.4 thousand in FY 1969 to 
$163.8 thousand in FY 1975 — a decrease of 25 per cent. 
Trends in Direct and Indirect Costs 

The IMPAC files contain separate records of funds expended for direct and 
indirect costs from all awards — for both NIH and ADAMHA. Based upon the total 
dollar amounts awarded to the entire sample of 145 universities, percentages 
have been computed for the two types of costs, by funding mechanism and fiscal 
year. The results are presented in Table 34 (for NIH) and Table 35 (for ADAMHA) 
on the following two pages . 



-81- 



Table 33 . Trends in Mean ADAMHA Awards to 145 Universities by Funding 



Mechanism- 


— Constant 


Dollars 


in Thousands 


(NIH R&D 


Deflator 




FY 1964 = 


100 ) a 


















Funding 










Fiscal Year 






Mechanism 


1969 




1970 




1971 


1972 


1973 


1974 


1975 


Regular Research Grants 


$ 78.2 


$ 


68.4 


$ 


68.7 


$ 71.2 


$ 69.9 


$ 73.2 


$ 68.8 


Program Project Grants 


7.1 




8.7 




8.3 


7.7 


6.2 


9.6 


4.0 


Clinical Research Grants 


— 




— 




— 


— 


— 


— 





Training Grants 


96.9 




89.9 




83.7 


79.4 


66.6 


83.7 


64.4 


Faculty Awards 


5.5 




5.5 




4.9 


4.2 


3.3 


2.6 


2.2 


Fellowships 


24.0 




23.7 




15.1 


14.9 


4.9 


5.5 


14.6 


Other Awards 


7.6 




9.1 




9.6 


9.5 


6.9 


9.5 


9.8 


All Categories 


219.4 


205.2 




L90.3 


186.9 


157.8 


184.1 


163.8 



Source: National Institutes of Health, IMPAC files. 



Table 33A . Trends in Index Numbers for the Means of ADAMHA Awards Shown in 
Table 33 (Means for FY 19 72 = 100) 



Funding 

Mechanism 



1969 



1970 



Fiscal Year 



1971 



1972 



1973 



1974 



1975 



Regular Research Grants 
Program Project Grants 
Clinical Research Grants 
Training Grants 
Faculty Awards 
Fellowships 
Other Awards 
All Categories 



109.7 


95.9 


96.4 


100.0 


98.1 


102.8 


96.5 


92.9 


114.0 


107.8 


100.0 


80.7 


125.3 


52.2 


122.0 


113.1 


105.4 


100.0 


83.8 


105.3 


81.0 


132.7 


131.7 


117.9 


100.0 


79.1 


63.0 


53.1 


160.7 


158.5 


100.7 


100.0 


32.7 


36.5 


97.6 


79.7 


95.7 


101.1 


100.0 


72.5 


99.5 


103.6 


117.3 


109.7 


101.8 


100.0 


84.4 


98.4 


87.6 



-82- 



Table 34 . Trends in Direct and Indirect Costs as Percentages of Total NIH 
Awards to 145 Universities by Funding Mechanism 3 



Funding Fiscal Year 

Mechanism 1969 1970 1971 1972 1973 1974 1975 



Regular Research Grants 
Direct Cost 
Indirect Cost 

Program Project Grants 
Direct Cost 
Indirect Cost 

Clinical Research Grants 
Direct Cost 
Indirect Cost 

Training Grants 
Direct Cost 
Indirect Cost 

Faculty Awards 
Direct Cost 
Indirect Cost 

Fellowships 

Direct Cost 100.0 100.0 100.0 100.0 100.0 100.0 100.0 

Indirect Cost — — — — — — — 

Other Awards 
Direct Cost 
Indirect Cost 

All Categories 
Direct Cost 
Indirect Cost 

Source: National Institutes of Health, IMP AC files, 



78.3% 


76.5% 


75.6% 


74.4% 


73.2% 


72.5% 


73.5% 


21.7 


23.5 


24.4 


25.6 


26.8 


27.5 


26.5 


81.8 


82.2 


81.8 


80.0 


76.1 


76.7 


74.3 


18.2 


17.8 


18.2 


20.0 


23.9 


23.3 


25.7 


98.2 


98.7 


98.5 


93.0 


94.6 


95.2 


94.9 


1.8 


1.3 


1.5 


7.0 


5.4 


4.8 


5.1 


93.8 


93.9 


94.0 


94.1 


94.2 


94.2 


94.2 


6.2 


6.1 


6.0 


5.9 


5.8 


5.8 


5.8 


94.3 


94.9 


95.8 


97.4 


98.0 


98.3 


98.8 


5.7 


5.1 


4.2 


2.6 


2.0 


1.7 


1.2 



98.8 


97.0 


94.5 


85.5 


80.7 


81.3 


86.8 


1.2 


3.0 


5.5 


14.5 


19.3 


18.7 


13.2 


84.3 


83.0 


81.7 


79.8 


77.4 


77.6 


78.0 


15.7 


17.0 


18.3 


20.2 


22.6 


22.4 


22.0 



-83- 



Table 35 . Trends in Direct and Indirect Costs as Percentages of Total 
ADAMHA Awards to 145 Universities by Funding Mechanism 3 



Funding 
Mechanism 



1969 



Fiscal Year 



1970 



1971 



1972 



1973 



1974 



1975 



Regular Research Grants 
Direct Cost 
Indirect Cost 

Program Project Grants 
Direct Cost 
Indirect Cost 



76.8% 


75.2% 


75.0% 


73.7% 


72.6% 


72.6% 


72.3% 


23.2 


24.8 


25.0 


26.3 


27.4 


27.4 


27.7 


76.5 


72.4 


71.8 


70.2 


72.3 


70.3 


70.0 


23.5 


27.6 


28.2 


29.8 


27.7 


29.7 


30.0 



Clinical Research Grants 
Direct Cost 
Indirect Costs 



Training Grants 
Direct Cost 
Indirect Cost 



94.2 
5.8 



94.0 
6.0 



94.0 
6.0 



94.0 
6.0 



94.0 
6.0 



94.0 
6.0 



94.0 
6.0 



Faculty Awards 
Direct Cost 
Indirect Cost 



92.7 
7.3 



92.7 
7.3 



92.4 
7.6 



92.8 
7.2 



92, 

7, 



97.8 
7.2 



92.7 
7.3 



Fellowships 
Direct Cost 
Indirect Cost 



98.7 
1.3 



97.4 
2.6 



95, 
4, 



94.8 
5.2 



88, 
11, 



86.4 
13.6 



95.5 
4.5 



Other Awards 
Direct Cost 
Indirect Cost 



76.4 
23,6 



76.6 
23.4 



80.0 
20.0 



80.0 
20.0 



81, 
18. 



80.2 
19.8 



83.3 
16.7 



All Categories 
Direct Cost 
Indirect Cost 



87.2 
12.8 



86.4 
13.6 



85, 
14, 



84.6 
15,4 



82.9 
17.1 



83.3 
16.7 



83.8 
16.2 



Source: National Institutes of Health, IMP AC files. 



-84- 



Th e most significant category for the evaluation of trends in indirect 
costs is that of regular research grants'—partly because most of the NIH funds 
are in that category. The figures in Table 34 show that the percentage of 
research grants allocated for indirect costs increased from 21.7 per cent in 
FY 1969 to 27.5 per cent in FY 1974— declining to 26.5 per cent in FY 1975. 
The indirect-cost trend was quite similar for ADAMHA research grants (see 
Table 35): ranging from 23.2 per cent in FY 1969 to 27.7 per cent in FY 1975. 

The trend results are generally similar in the case of program project 
grants. For NIH, the indirect-cost proportions increased from 18.2 to 25.7 per 
cent over the seven-year period; while for ADAMHA the corresponding increase 
was from 23.5 per cent in FY 1969 to 30 per cent in FY 1975. 

(The determinants of the division of costs into the direct and indirect 
components are so variable for the remainder of the funding categories that 
discussion of the remainder of the results in Tables 34 and 35 hardly seems 
justified. ) 

There are several reasons why indirect costs have increased more rapidly 
than the direct costs of research in recent years. Probably the most signifi- 
cant has been the differential effects of inflation on the two types of R&D 
expenditures. The direct-cost component has a higher proportion of salaries 
and wages than the indirect-cost sector; and the former has been increasing at 
lower rates than the non-personnel elements of the latter. Price increases for 
utilities (especially fuel costs), books and periodicals, and other non-person- 
nel items have escalated sharply in recent years, whereas increases in staff 
compensation have been at far lower rates. Another factor has been the marked 



-85- 



increases in administrative costs (mostly classified as indirect costs) and 
other costs due to federally mandated social programs such as equal employment 
opportunity, occupational safety and health, enviornmental protection, and 
fair-labor standards. Compliance with regulations governing the use of human 
subjects and animals in experimentation adds especially to both the direct and 
indirect costs of biomedical-behavioral research. And the increasingly detailed 
information requirements under grants and contracts have added substantially 
to the indirect costs of sponsored research. 



-86- 



IX. SUMMARY AND CONCLUSIONS 

The primary purpose of this study was to investigate relationships between 
trends in federal funding of biomedical-behavioral research in universities and 
concurrent financial and educational changes in these institutions from FY 1964 
through FY 1974. It was assumed that more detailed knowledge of these inter- 
relationships would provide guidance to federal agencies and to universities 
in their interdependent efforts to sustain academic research in the health fields 
at a high level of national ef fectivenss. 

The analytical framework for the investigation involved two modes of trend 
comparisons: (a) differences among various types of institutions in patterns 
of change in a given financial or educational variable; (b) concurrent changes 
in two or more variables for a given type of institution. 

The sample of 148 universities had a "three-dimensional" structure: (a) 
type of control (public, private); (b) medical-school status (presence, absence); 
(c) classification in terms of involvement in doctoral education and research. 
Fifty-five of the university campuses were private and 80 were public in 
governance. Sixty-eight had medical schools under their campus jurisdiction. 
The entire sample comprised all but nine of the 145 institutions in the top 
three categories of the classification system developed by the Carnegie Commis- 
sion on Higher Education. (The nine omissions from the latter were due to large 
amounts of missing data.) Twelve institutions from other Carnegie Commission 
categories were included because eight of them had medical schools and the other 
four were members of multi-campus institutions whose aggregate data seemed 
likely to be useful in certain of the anticipated analyses. 



-87- 



The statistical information used in the study included several types of 
financial and nonfinancial data spanning the period from FY 1964 through FY 1974. 
More than 150 data elements were sought from the files of four federal and two 
private agencies, covering each of the 11 years and all institutions in the 
sample. Unfortunately, the quantity of data available fell short of the desired 
amount for several reasons: (a) most of the surveys did not span the entire 11- 
year period and some that did were conducted intermittently; (b) changes in the 
definitions of data elements occurred in certain surveys; (c) many institutions 
failed to supply data for one or more years in a series; (d) numerous changes 
occurred in the identity of the reporting unit, mainly in the case of multi- 
campus institutions. Moreover, the quality of the data varied considerably 
among agencies and sometimes among the data files of a single agency. 

Summary of Findings 

Despite the limitations of the data base, it was possible to complete a 
wide range of trend analyses which disclosed important differences among the 
variables by type of institution. 

Trends in educational-and-general (E&G) revenues . The importance of dis- 
aggregating the results for the composite sample of research universities into 
relatively homogenous sub-groups was strongly demonstrated by the analysis of 
the data for E&G revenues. 

1. For the entire group of 100 institutions, E&G revenues in constant 
dollars increased slowly every year from FY 1969 through 1974 (except FY 
1973 when there was essentially no change). The overall gain was 11.1 
per cent. 

2. Dividing the sample into private and public universities showed gains 
of 2.0 per cent for the private and 19.7 per cent for the public univer- 
sities over the seven years. 



-88- 



3. Particularly striking were the differences between the private and 
public institutions from FY 1972 through FY 1974: private E&G revenues 
declined 3.1 per cent, while public E&G revenues gained 7.1 per cent. 

4. Analysis of E&G revenues for the private institutions by the categories 
of the Carnegie Commission showed that Research Universities I accounted 
for all of the decrease for the private sector from FY 1972 to 1974. 
Research Universities II and "Other Categories" had increases of 2.5 and 
6.5 per cent, respectively. 

The proportion of E&G revenues supplied by R&D funds . Both total and 
federally funded R&D revenues for the sample of 100 universities declined mod- 
erately as proportions of E&G revenues, from FY 1969 through 1974: from 21 to 
19 per cent for total R&D funds, and from 18 to 15 per cent for federal R&D 
funds. From these figures it can be calculated that federal R&D funds dropped 
from 86 to 79 per cent of total R&D revenues during the six years. Other speci- 
fic findings: 

1. R&D funds constituted a much higher percentage of the E&G funds of 
private than of public universities (e.g., 28 vs. 16 per cent in FY 1969 
for total R&D, and 24 vs. 13 per cent for federal R&D funds). 

2. Institutions with medical schools had higher proportions of R&D funds 
than those without medical schools (e.g., 24 vs. 19 per cent for total 
R&D revenues, and 20 vs. 16 per cent for federal R&D funds in FY 1969). 

Trends in sponsored R&D revenues for all fields . Consistent with the 
findings just discussed, the mean amount per institution of total and of federal 
R&D funds in constant dollars declined over the six-year period for all insti- 
tutions combined (6.1 per cent for total and 10.2 per cent for federal R&D funds). 

1. Private institutions accounted for all of the decline in R&D funds, 
however, as with E&G funds discussed earlier. From FY 1969 to 1974, private 
institutions suffered declines of 12 per cent in total R&D funds and of 
17 per cent in federal R&D funds. Public institutions, on the other hand, 
gained about five per cent in total R&D funds and less than one per cent 
in federal funds. 



-89- 



2. Most of the decline in R&D funding for private institutions occurred 
in those without medical schools. But for public institutions, their 
small increases were associated with absence of medical schools. 

3. Unlike the trends found for E&G revenues, private universities showed 
no appreciable differences when grouped into the three Carnegie Commission 
categories. 

Expenditures for biomedical-behavioral research . Results are summarized 

first for R&D expenditures in all biomedical-behavioral fields combined and then 

for three of the disciplines recognized in NSF's expenditures survey (see 

Chapter V) : 

1. Total R&D funding for all biomedical-behavioral sciences combined showed 
a marked upturn of 10.6 per cent in FY 1973 (followed by a slight decline 
from that level in FY 1974) , in sharp contrast to trends in other fields. 
Both private and public universities showed similar patterns. Institutions 
without medical schools fared better generally than those with medical 
schools. 

2. The trends for federally funded biomedical-behavioral expenditures 
were similar to those for all R&D funding in these fields; but the federal 
component showed a sharper upturn of over 12 per cent in FY 1973 (for pri- 
vate and public universities combined) , followed by a decline in FY 1974 
averaging about 5 per cent. Most of the sub-groups of institutions showed 
parallel trends. 

3. Among the disciplines within the biomedical-behavioral complex, the 
most marked upturn in federal funding in FY 1973 for all institutions com- 
bined occurred for the biological sciences (22 per cent) ; but more than 
half of that gain was lost in the following year. 

Federal funding for medical sciences increased moderately in FY 1973 (6 
per cent), but showed a further rise in FY 1974 to 10 per cent above the 
FY 1972 level. 

Psychology had a slight increase in federal R&D funds of about 3 per cent 
in FY 1973, but then a drop in 1974 of almost 5 per cent below the FY 1972 
level. 



-90- 



An analysis of NIH awards for basic research grants to 145 universities in 
the sample showed an overall increase from FY 1969 through 1975 of 35 per cent. 
Although these obligation figures are not strictly in phase with NSF's biomedi- 
cal expenditure data, it is interesting to note that the latter T s total increase 
over the six years was 40 per cent for federally funded R&D. 

A seven-year increase of 35 per cent in NIH funds awarded for regular 
research grants was paralleled by an increase of 22 per cent in the proportion 
of the grant totals allocated for indirect costs. Thus the constant-dollar 
increase in funds available for the direct costs of research was reduced to 27 
per cent. 

Trends in enrollment . Three types of fall-term enrollment statistics were 
analyzed: total degree-credit enrollment; enrollment for advanced degrees in 
all fields combined; enrollment for advanced degrees in biomedical-behavioral 
sciences. 

1. Total degree-credit enrollment . For all institutions combined, total 
enrollment exhibited fairly steady growth from the fall term of 1965 through 
1971, with a further increase of only two per cent through FY 1974. E&G 
revenue increases from FY 1969 through 1974 showed a growth of 11.1 per 
cent and a parallel growth of 13.6 per cent in enrollment occurred for the 
same six-year period. 

The enrollment trends for private and public universities differed consi- 
derably: private institutions showed almost no change over the six-year 
period, while enrollment increased about 18 per cent for the public uni- 
versities. Within each group, there were only slight differences among 
the three Carnegie Commission categories. 

2. Enrollment for advanced degrees in all fields (not including profes - 
sional degrees in medicine, etc .). The general growth patterns of enroll- 
ment for graduate degrees for all institutions combined were similar to 
those for total degree-credit enrollment. (No data for advanced-degree 
enrollment were available for FY 1974.) Growth was fairly steady from the 
fall term of 1967 through 1971, leveling off thereafter. 



-91- 



Private institutions tended to maintain parity with public institutions 
in enrollment growth at the graduate level; and presence or absence of a 
medical school made little difference. Moreover, there were no signifi- 
cant differences among the Carnegie Commission categories in advanced-degree 
enrollment. 

While graduate enrollment increased about 10 per cent from FY 1969 to 1973, 
R&D revenues declined (2 per cent for all sources of funds and about 7 
per cent for federal funds) . 

3. Enrollment for advanced degrees in biomedical-behavioral sciences . As 
compared with enrollment for advanced degrees in all fields, the rate of 
growth was substantially higher for the biomedical-behavioral sciences. 
Growth did not level off in 1971, but instead continued through the fall 
term of 1973. Private institutions showed slightly higher rates of growth 
than their public counterparts after 1970: 8.7 per cent in 1973 over 1972 
vs. 3.6 per cent for public universities. 

From FY 1968 through 1973, biomedical-behavioral enrollment increased more 
rapidly than federal R&D funding in these fields (30 per cent vs. 11 per 
cent). But between FY 1972 and 1973, the corresponding increases were 5 
and 12 per cent. 

Trends in doctoral degrees (not including M.D.'s, D.D.S.'s, etc.) . In view 

of the irregular relationship between number of doctorates granted and financial 

data for a given year, no comparisons of trend indices for the two sets of 

variables will be included here. 

1. Doctoral degrees in all fields . There was fairly steady growth in the 
mean number of doctorates granted by all institutions combined from FY 1964 
through FY 1973, with a leveling off or slight drop in FY 1974. The trend 
for private universities was similar to that for the public sample. But 
when institutions within these two groups were classified by presence or 
absence of a medical school, the private and public groups showed somewhat 
opposite trends: private institutions with medical schools showed a lower 
growth rate (and a decline in 1974) ; whereas the public institutions with 
medical schools had somewhat higher average increases than those without 
medical schools over most of the 11-year period. 

2. Doctoral degrees in biomedical-behavioral sciences . For all institu- 
tions combined, the mean index was higher before FY 1971 relative to the 
base year (FY 1972) for biomedical-behavioral doctorates than for all 
fields combined; and the former's growth after FY 1970 was at a somewhat 
higher rate than the latter* s. But the differential favoring the bio- 
medical-behavioral fields was not as great as the parallel comparison 
involving total enrollment for advanced degrees. 



-92- 

The private universities showed substantially higher growth rates than 
public universities for bioniedical-behavioral doctorates after FY 1971. 
For the public institutions, these degrees levelled off after FY 1971. 
Institutions with medical schools showed increases after FY 1972, while 
those without them declined slightly. 

General Conclusions 

The findings of this study have not provided, unfortunately, a conclusive 
answer to the critical question of whether the nation's research universities 
can continue to sustain an adequate level of health-related research with their 
present resources and financial prospects. This indeterminancy stems partly 
from the terminal year of FY 1974 for the data base, which meant that the effects 
of the downturn in the national economy during that year and since could not be 
fully reflected in the financial analysis. But deficiencies in the quantity 
and quality of the information available were also important barriers to a more 
definitive determination of that issue. 

Nevertheless, several key findings seem to justify the conclusion that the 
financial condition of private research universities — especially those described 
as Research Universities I in the Carnegie Commission's classification — has been 
deteriorating in recent years under the joint impacts of cost inflation and 
recession in revenues. 

Although the public universities through FY 1974 appear to have maintained 
a stable or slightly increased constant-dollar level of revenues, the latter 
did not quite keep pace with enrollment increases. Furthermore, an earlier 
ACE study based on expenditures per full-time-equivalent student showed that 
when the constant-dollar analysis was extended through FY 1975, all public insti- 
tutions showed declines and all categories of private institutions showed 
further erosion from the level to which they had been reduced in FY 1974. It 
seems probable that this negative trend has continued into the current year in 
both sectors. 



-93- 



Contributing to these financial difficulties has been the decline in federal 
R&D funds as proportions of total educational budgets — more severe for private 
than for public universities and especially sharp for both in FY 1974. Since the 
growing financial pressures upon institutional budgets as a whole would not 
allow universities generally to use other revenues to offset declines in federal 
R&D funding, it seems reasonable to conclude that the overall R&D base of research 
universities may have suffered serious impairment. 

The one area that showed a positive R&D funding trend in this study was the 
biomedical-behavioral sciences. They had moderate revenue increases in constant 
dollars, especially between FY 1972 and 1974, in both public and private univer- 
sities. Whether or not this trend has continued is unknown. Nor is there direct 
evidence bearing upon the impact of the funding growth in these fields upon the 
financial and educational resources of the universities as a whole. It is a 
plausible hypothesis, however, that unilateral R&D growth in the health fields — 
accompanied, as has occurred, by substantial expansion in training programs for 
the health professions — probably has put strong pressure upon university budgets 
already under stresses caused by inflation/recession. 

An important contributor to this burden has been the requirement of sharing 
in the cost of academic research supported by federal grants. Limitations 
upon the full recovery of the indirect costs of federally sponsored research 
has created a related drain upon university resources at a time of rapid price 
escalation in such costs — much of the latter due to federally mandated social 
programs and to compliance with other kinds of federal regulations. 



-94- 



Th e effects of inflation on all costs of research, together with the dis- 
proportionate increase in the indirect-cost components, have serious implica- 
tions for academic science departments and federal program directors who share 
an interest in maximizing the amount and quality of research produced with 
available funds. The inevitable result of expending fixed amounts of grant 
funds to meet such costs is to lower research "productivity" as measured by the 
proportionate level of investment in direct scientific effort. It is natural 
for department heads and federal program directors to try to avoid such an out- 
come, which is likely to mean putting pressure upon institutions to accept 
grants with as much cost-sharing as can be negotiated. Although understandable, 
this kind of practice cannot be continued without aggravation of the serious 
financial difficulties facing research universities. Among these problems are 
the generation of undesirable budgetary conflicts among departments and compe- 
tition for funds needed for such campus-wide purposes as faculty compensation, 
libraries, and plant maintenance. 

These and other problems arising from the interrelations between federal 
research agencies and universities need to be more intensively studied in order 
to establish a sounder foundation for long-range policies and programs fostering 
their mutual interests. The kind of "macroscopic" information used in the 
present study can at best discover broad trends, identify puzzling problems, 
and suggest speculative hypotheses. Far more detailed knowledge of intra- 
institutional "dynamics" is necessary to an adequate understanding of the 
"impact" upon university finances and programs of such federal policies as cost 
sharing, sudden changes in funding levels, and major shifts in program priorities. 



-95- 



During the 1960s, in an expanding educational economy, universities could 
usually manage to adapt to such changes without undue disruption of their 
educational programs or damage to their financial stability. This is no longer 
the case. And if universities are to continue to perform their distinctive 
role in the national research effort effectively, new ways and means must be 
found to encourage more productive interaction between the federal government 
and academic institutions in pursuit of their common research goals. 



-96- 



REFERENCES 



1. Bowen, William G. The Effects of Inflation/Recession on Higher Educa- 

tion. Educational Record , Summer 1975, Volume 56, No. 3. 

2. Carnegie Commission on Higher Education. A Classification of Insti - 

tutions of Higher Education . A Technical Report. Berkeley, 
Calif. : Carnegie Commission on Higher Education, 1973. 

3. Drew, David E. , and Wirt, John G. The Effects of Federal Funds upon 

Selected Health-Related Disciplines . The Rand Corporation, 
R-1944-PBRP, March 1976. 

4. Halstead, D. Kent. Higher Education Prices and Price Indexes . Washing- 

ton, D.C.: U.S. Government Printing Office, 1975. 

5. Lanier, Lyle H. , and Andersen, Charles J. A Study of the Financial 

Condition of Colleges and Universities: 1972-1975 . An ACE 
Special Report. Washington, D.C.: American Council on Education, 
October 1975. 

6. Morgan, T.E., and Jones, D.D. Trends and Dimensions of Biomedical 

and Behavioral Research Funding in Academic Medical Centers : 
1964-1974 . Association of American Medical Colleges, February 
1976. 

7. Science Indicators, 1974 . The National Science Board/National Science 

Foundation, 1975. 

8. Williams, A. P., Carter, G.M. , Chu . D.S.C., Coleman, S. , Massell, A. P., 

Neu, C.R., Rasmussen, R. , and Rogers, W. The Effect of Federal 
Biomedical Research Programs on Academic Medical Centers . The 
Rand Corporation, R-1943-PBRP, March 1976. 

9. Woodrow, Raymond J. Indirect Costs in Universities . An ACE Special 

Report. Washington, D.C.: American Council on Education, March 
1976. 



-97- 

APPENDIX A 

CLASSIFICATION OF THE 148 INSTITUTIONS IN THE ACE-RAND SAMPLE B 
THE CATEGORIES OF THE CARNEGIE COMMISSION ON HIGHER EDUCATION 5 



% 



PRIVATE INSTITUTIONS 



Research Universities I 



California Institute of Technology 
Case Western Reserve University (M) 
Columbia University, Main Division (M) 
Cornell University (M) 
Duke University (M) 
Harvard University (M) 
Johns Hopkins University (M) 
Massachusetts Institute of Technology 
New York University (M) 
Northwestern University (M) 
Princeton University 

Research Universities II 



Rockefeller University 
Stanford University (M) 
University of Chicago (M) 
University of Miami (M) 
University of Pennsylvania (M) 
University of Rochester (M) 
University of Southern California (M) 
Vanderbilt University (M) 
Washington University (M) 
Yale University (M) 
Yeshiva University (M) 



Brandeis University 

Brown University (M) 

Boston University (M) 

Carnegie-Mellon University 

Catholic University of America 

Claremont Graduate School 

Emory University, Main Campus (M) 

Doctoral Universities I 

American University 
Boston College 
Brigham Young University 
Dartmouth College (M) 
Fordham University 
Georgetown University (M) 
Howard University (M) 
Lehigh University 
Loyola University (M) 

Other Categories 

Creighton University (M) 
Loma Linda University (M) 



George Washington University (M) 

Illinois Institute of Technology 

Rice University 

Syracuse University 

Tufts University (M) 

Tulane University of Louisiana (M) 



Marquette University 

Northeastern University 

Pennsylvania Drexel University 

Rensselaer Polytechnic 

St. John's University 

St. Louis University, Main Campus (M) 

University of Denver 

University of Notre Dame 



Wake Forest University (M) 



Carnegie Commission on Higher Education. A Classification of Institutions of 
Higher Education. Berkeley, Calif. : Carnegie Commission on Higher Education, 1973. 

The institutions listed are generally on single campuses, and those with medical 
schools under campus jurisdiction are designated by "(M)". 



A-l 



Research Universities I 



-98- 



PUBLIC INSTITUTIONS 



Michigan State University (M) Univers 

North Carolina State Univ. , Raleigh Univers 

Ohio State University, Main Campus (M) Univers 

Purdue University, Main Campus Univers 

Rutgers University, New Brunswick (M) Univ. o 

Texas Agricultural & Mechanical Univ. Univ. o 

University of Arizona (M) Univ. o 

University of California, Berkeley Univ. o 

University of California, Davis (M) Univers 

Univ. of California, Los Angeles (M) Univers 

University of California, San Diego (M) Univers 

University of Colorado (M) Univers 

University of Florida (M) Univers 
University of Georgia 



ity of Hawaii, Main Campus (M) 

ity of Illinois, Urbana 

ity of Iowa (M) 

ity of Kentucky, Main Campus (M) 

f Maryland, College Park 

f Michigan, Main Campus (M) 

f North Carolina, Chapel Hill (M) 

f Pittsburgh, Main Campus (M) 

ity of Tennessee, Knoxville 

ity of Texas, Austin 

ity of Utah (M) 

ity of Washington (M) 

ity of Wisconsin, Madison (M) 



Research Universities II 



Auburn University, Main Campus 
City Univ. of N.Y., Graduate Center 
Colorado State University 
Florida State University 
Georgia Institute of Technology 
Indiana University, Bloomington 
Iowa State Univ. of Science & Tech. 
Kansas State Univ. of Agriculture 

and Applied Science 
Louisiana State Univ. , Baton Rouge 
Mississippi State University 
Oklahoma State University 
Oregon State University 
State Univ. of N.Y., Buffalo 



Temple University, Main Campus (M) 

Univ. of Arkansas, Main Campus 

Univ. of Cincinnati, Main Campus (M) 

Univ. of Connecticut, All (M) 

University of Kansas (M) 

Univ. of Massachusetts, Amherst 

Univ. of Nebraska, Lincoln 

Univ. of Oklahoma, (M) 

University of Oregon, Main Campus 

University of Virginia, All (M) 

Virginia Polytechnic Inst. & State Univ. 

Washington State University 

Wayne State University (M) 

West Virginia University (M) 



Doctoral Universities I 



Arizona State University 

Ball State University 

Clemson University, 

Kent State University, Main Campus 

Montana State University 

New Mexico State Univ. , Main Campus 

North Dakota State Univ., Main Campus 

Ohio University, Main Campus 

Southern Illinois Univ., Main Campus (M) 

State Univ. of New York, Albany 

State Univ. of N.Y., Stony Brook 

University of Alabama, Main Campus 

University of California, Riverside 

University of California, Santa Barbara 

University of Delaware 



University of Houston, Main Campus 
University of Idaho 
University of Louisville (M) 
University of Maine, Orono 
University of Mississippi (M) 
University of New Hampshire 
University of New Mexico, Main Campus 
University of North Dakota, Main Campus 
University of Rhode Island 
Univ. of South Carolina, Main Campus 
University of South Dakota, Main Campus- 
University of Southern Mississippi 
Univ. of Vermont & State Agric. College 
University of Wyoming 
Utah State University 



(M) 
(M) 



A-2 



-99- 



PUBLIC INSTITUTIONS (cont.) 



Other Categories 



Texas Technological University (M) University of South Florida (M) 

University of Alabama, Birmingham (M) University of Texas, Arlington 

University of Illinois, Chicago Circle University of Wisconsin, Milwaukee 

University of Nevada, Reno (M) Virginia Commonwealth University (M) 
Univ. of Puerto Rico, Rio Piedras 



= e 



A-3 



-100- 



AP-PENDIX B 



TABLES SHOWING EXPENDITURES FOR BIOMEDICAL-BEHAVIORAL 

RESEARCH FOR INSTITUTIONS CLASSIFIED BY CATEGORIES 

OF THE CARNEGIE COMMISSION ON HIGHER EDUCATION 



TABLE B-l. 



TABLE B-1A. 



Trends in Mean Expenditures for All Biomedical-Behavioral Research 
by Carnegie Commission Categories of Institutions — Constant Dollars 
in Thousands (NIH R&D Deflator, FY 1964 = 100) 

Trends in Index Numbers for the Means of All Biomedical-Behav- 
ioral Research Expenditures Shown in TABLE B-l (Means for FY 
1972 = 100) 



TABLE B-2. Trends in Mean Federally Funded Expenditures for Biomedical- 
Behavioral Research by Carnegie Commission Categories of Insti- 
tutions — Constant Dollars in Thousands (NIH R&D Deflator, FY 
1964 = 100) 

TABLE B-2A. Trends in Index Numbers for the Means of Federally Funded Biomedi- 
cal-Behavioral Research Expenditures Shown in TABLE B-2 (Means 
for 1972 = 100) 



TABLE B-3. 



TABLE B-3A. 



Trends in Mean Federally Funded Expenditures for Biological Re- 
search by Carnegie Commission Categories of Institutions — Constant 
Dollars in Thousands (NIH R&D Deflator, FY 1964 = 100) 

Trends in Index Numbers for the Means of Federally Funded Expen- 
ditures for Biological Research Shown in TABLE B-3 (Means for 
FY 1972 = 100) 



TABLE B-4. Trends in Mean Federally Funded Expenditures for Medical Research 

by Carnegie Commission Categories of Institutions — Constant Dollars 
in Thousands (NIH R&D Deflator, FY 1964 = 100) 

TABLE B-4A. Trends in Index Numbers for the Means of Federally Funded Expen- 
ditures for Medical Research Shown in TABLE B-4 (Means for FY 
1972 = 100) 



TABLE B-5. 



TABLE B-5A. 



Trends in Mean Federally Funded Expenditures for Research in 
Life Sciences not Elsewhere Classified, by Carnegie Commission 
Categories of Institutions — Constant Dollars in Thousands (NIH 
R&D Deflator, FY 1964 = 100) 

Trends in Index Numbers for the Means of Expenditures for Research 
in Life Sciences not Elsewhere Classified Shown in TABLE B-5 (Means 
for FY 1972 = 100) 



TABLE B-6. Trends in Mean Federally Funded Expenditures for Research in 

Psychology by Carnegie Commission Categories of Institutions — 
Constant Dollars in Thousands (NIH R&D Deflator, FY 1964 = 100) 

TABLE B-6A. Trends in Index Numbers for Means of Federally Funded Expendi- 
tures for Research in Psychology Shown in TABLE B-6 (Means for 
FY 1972 = 100) 



B-l 



-101- 



PN 


rD 






C 


,G 


•H 


<J 




U 


CO 


re 


M 


a 


re 


01 


rt 


CJ 


rH 


pes 


O 




O 


t-\ 




re 


4J 


u 


c 





CO 


■H 


4-1 


> 


01 


re 


c 


j= 





01 





CO 
1 


1 

1 


<H 


CO 


re 


C 





O n) 


■H 


t-t >-n. 


"0 


4J O 


aj 


3 O 


E 


4-1 rH 





•H 


T-l 


■U II 


pg 


co 




C -a- 


rH 


M CO 


rH 


on 


< 


<4-l rH 




O 


u 




CO C 


m 


01 




•H « 


CO 


>-. M 


al 


O O 


M 


60 4J 


3 


CU fl> 


4J 


4-1 i-l 


•H 


m <*-» 


T3 


cj cu 


C 


a 


CD 


a 


a. 





x 


■h »a 


w 


co pd 




CO 


c 
re 


@3 


01 


B ES 


a 


nj 




CJ 


c 


co 


■H 


CU 13 


to 


em re 


T3 


cu co 


C 


a a 


01 


n 


H 


«3g 


rH 
1 
CQ 




W 




rj 




S 




H 





1-4 

CU 

a 
z 



p 

o 

4-1 -H 
O 4-1 

3 

CU 4J 
O. -H 
r>N 4J 
H 0) 

c 



O CM 

ON -J- 

co m 


in 

CO 

ro 


00 
00 


O 

-3- 

CO 


in 

ON 


CM 
MN 


IN 

in 
















VO co 
-H 


CM 




m 


.H 
rH 


Mf 


rH 


m m 
nT on 

m in 


O 

in 


00 
on 


CO 

00 
co 


vO 

CO 

in 


CO 


m 

m 
















*c co 

rH 

co- 


CM 




m 


rH 
rH 


<r 


rH 


co cm 
co 
O 00 


CM 
rH 


CO 

mi- 
co 


-3- 

in 


CM 

O 

m 


m 

in 
O 


cn 

CM 
CM 
















\C CM 

■H 
(ft- 


CJ 




-3- 




rH 


<f 


rH 


co m 

ON vO 

vc m 


in 

CNI 

m 


CM 




co 



-H 
O 

ro 


rH 


O 
CO 

rH 


rH 


CNl 


rH 


m 


rH 
rH 


n3 


rH 


<t 
on rN 

14B co 


m 


V43 

in 

CM 


CO 

-cf 


00 
ON 

m 


ON 
O 


0O 

CO 




in rH 


CM 


rH 


-4T 


ON 


-<r 


rH 


m 00 

vO V43 
cm co 



in 

rH 


00 

-a- 
cri 


OS 

O 

00 


MD 

in 

ON 




CM 

m 


rH 
CO 
O 
















st 00 


CM 




CO 


n. 


CO 


rH 


v> 














<t CM 

in cm 


ro 


on 

rH 


00 


-£1 

CM 


in 

CM 


CO 



CO 
















c 


CO 


CO 




CO 


CO 


CO 







cu 


CU 




c 


CU 


cu 




•H 


•H 


•H 







•H 


•r-i 




4J 


4-1 


4-1 




•H 


4J 


4-1 




3 


•H 


•H 




4J 


•H 


•H 




4J 


CO 


to 


CO 


3 


CQ 


CO 


CO 


•H 


t-i 


l-l 


cu 


4-1 


tJ 


S-i 


0) 


4-1 


01 


CU 


•H 


•H 


CU 


CU 


•H 


CO 


> 


> 


1-1 


4J 


> 


!> 


l-l 


c 


•H 


-H 





co 


•H 


•H 





M 


c 


c 


00 


c 


c 


C 


00 




3 


3 


CU 


M 


3 


O 


cu 


01 






4J 








4-> 


4-1 


-C 


J= 


CO 


O 


rC 


-C 


(0 


re 


u 


O 


CJ 


•H 


O 


CJ 


CJ 


> 


S-i 


V-i 




rH 


rJ 


M 




-H 


re 


re 


r<4 


J3 


re 


re 


u 


U 


CU 


0) 


CU 


3 


cu 


0) 


01 


P* 


CO 


CO 


.C 


Pi 


CO 


CO 


£ 




0) 


0) 


4-1 




QJ 


cu 


4-1 


r-< 


K 


OS 





rH 


OS 


OC 






CO 

-o 

c 

3 

o 

u. 

<u 

CJ 

c 
cu 

•H 
U 
CO 



re 
c 
o 



n 



cu 

CJ 

u 
3 
o 
in 



re 


p 


IH 


O 


n 


rH 


•H 




> 


II 


re 




j= 


CM 


OJ 


r^ 


CQ 


ON 


1 


-H 



CO >, 

CJ [t, 
•H 
T3 U 

cu o 

e uh 
o 

•H CO 

CQ C 

CO 

rH CU 

< -^ 

•4-1 rH 

O I 
CO 
CO 

C W 
CO _] 

^5 
H 
cu 

j: a 



u a 

o 3 

iw O 

co co 

H 

CU CO 

XI OJ 

a u 

3 3 

55 4J 

•H 

!< -O 

CU 13 

TJ cu 

C O. 

M X 

w 



TJ 

c 



r-l| 



ON 00 ^3 14O 



m m o o 
O O rH o 



<■ vO or^ 



rH r^ m *£> 

HOHCN 



rM 

CU 

1 

3 
S3 



c 

o 

14-1 -H 
O 4-1 
3 
CU 4-1 
O. -H 
>N 4-1 
H CO 

c 



moorN.oN n co -jr> 



co in in co 
o O rH o 



0000 



000 
000 



-3- CO CO O 



ON CTi rH CM 



-3 CO CO «} 



-I CONH 
ON CO rH in 



r~ -a- mi- -j- 



o m o cn 

IN ^O O rH 



N3" CN CO ON 
in CM rH rH 



CM rH CJN CM 

HHON 



OOOO 

o o 00 

OOOO 

rH rH rH rH 



O ^O CM CM 



in IN CM lO 

OOOOl 



00 -a- cn in 



vO rH COCO 
ON CTi rH 00 



m co 00 



on m *& <r 
CN h. coco 



On *£> IN V.O 
00 CM CM CO 



CO 
















c 


CO 


co 




CO 


CO 


CO 







OJ 


cu 




c 


OJ 


cu 




■H 


■H 


•H 







•H 


•H 




4J 


l-l 


4J 




•H 


4J 


4J 




3 


•H 


■H 




4J 


■H 


•H 




4J 


CO 


CO 


CO 


3 


CO 


ca 


•j> 


•H 


u 


u 


cu 


4J 


rX 


r^ 


01 


4J 


0) 


V 


•H 


•H 


OJ 


0) 


•H 


CO 


> 


> 


rJ 


4-1 


> 


> 


1-4 


c 


•H 


•H 


O 


CO 


■H 


■H 





rH 


c 


C 


60 


c 


C 





00 




3 


s 


cu 


IH 


O 


H 


01 


0) 






4J 








4-1 


4-1 


X 


-C 


CO 


a 


rC 


J^ 


re 


CO 


CJ 


CJ 


CJ 


•H 


O 





O 


> 


S-i 


u 




rH 


u 


ki 




•H 


re 


re 


U 


XI 


re 


re 


1-1 


l-l 


cu 


0) 


cu 


3 


0) 


0) 


0) 


P-, 


CO 


CO 


j= 


Cm 


CO 


CO 


rC 




0) 


01 


4-1 




OJ 


cu 


4J 


^H 


« 


OS 





rH 


os 


£ 


O 



B-? 



-102- 



H 


4J 







C 




■H 


re 




> 


4-J 




rt 


a 




£ 


c 




01 


o 




00 


o 




rH 


1 




cfl 


CO 




U 


c 




•H 


o 




-o 


1-1 


ra 


cu 


4-1 


^■N 


a 


3 


o 


G 


4-J 


o 


■H 


•H 


rH 


to 


4-> 






0) 


II 


M 


c 







rH 


<r 


14H 




mo 




<H 


on 


0) 





.h 


<u 






r-l 

3 


en 

01 


~£ 


4-1 


•H 




■H 


ij 


* 


-a 


o 


h 


C 


10 


o 


01 


01 


u 


a. 


JH 


ra 


X 


ra 


rH 


w 


c; 


0J 


13 


c 


a 


0) 


o 




T3 


•H 


a 


C 


C0 


-a 


3 


cn 


os 


fa 


■H 


X 


>, 


£ 


H 


H 





Z 


H 


o 


* — ' 


ra 






H 


01 


0) 


01 


•h 


T3 


TO 


00 


C 


0) 


0) 


ra 


fa 


£ 


cn 

3 


C 


ra 


O 


0> 


e_> 


£ 


s 


>. 






jo 


c 


C 




■H 


■H 


jo 






cj 


cn 


cn 


1-4 


i- 


-o 


m 


ra 


c 


01 


rH 


0) 


CO 


•h 


u 


01 


o 


H 


PS 


n 



0) 

•i 

3 
Z 



B 
o 

IH -H 

4J 

3 

01 4-> 
Q. -H 
>N 4J 

£-• cn 
C 



co\omtM 
is -a- m ^ 

O N CJivO 



in eg <r vo 

•j n ism 
O in O rs 



<f oo is o 
-3- m in oo 

o 00 CN 00 



o mo m is 

■-I cn io cm 
cn] m i-H oo 



cn cn 


rH 


mO 


in 


U0 


rH 


cn 




CN IS 


m 


00 


n 


CM 


<r 


(7\ 




in s>- 


rs 


M3 


CO 


^O 


o 


iO 




A M 


A 




a 


•^ 


A 






<r a> 


rH 




CM 


■~o 


CN 






</> 
















o on 


MO 


O 


CO 


cn 


-H 


■H 




CN i-H 


H 


-<r 


Sf 


o 


cn 


CT\ 




m tH 


rH 


oo 


rH 


m 


CM 


vD 




















<f ON 


CN 




m 


rs 


CN 






<J> 
















i-l CN 


.-4. 


CTN 


rs 


is 


ON 


O 




■H tH 


tn 


rH 


o 


m 


rH 


cn 




is .J- 


i-H 


o 


ON 


CM 


rs 


M3 




















<■ ON 


CN 


■H 


CN 


MT' 


CM 






■Ho- 
















rn m 


<r 


CO 


m 


m 


M3 


vO 




k£> vO 


CN 


vD 


CO 


CO 


ON 


C3> 




sj- is 


CO 


is 


rH 


^o 


CO 


m 




*■ A 
















cn mo 


H 




CN 


nT 


rH 




c 




</> 














■H 
4-1 

ra 


sf CN 


cn 


ON 


ON 


■JD 


rs 


lO 


-a 


m cn 


rH 
1-4 


.-1 


CO 


CM 


CN 

M 


cn 


c 

3 

o 


M 


H 






rH 


rH 




o 


10 














c 


c cn 


cn 




cn 


cn 


cn 




CJ 


0) 


oi 




c 


0) 


01 




-H 


•H -H 


•H 




o 


•H 


'H 




o 


4-1 4-1 


4-1 




•H 


4-1 


4-1 




c-j 


3 -H 


■H 




4-1 


•H 


•H 






4-> cn 


cn 


CD 


3 


CO 


cn 


CO 


rH 


•H M 


!H 


0J 


U 


1-4 


i-i 


01 


ra 


J-l a) 


01 


•H 


•H 


0J 


0) 


•H 


c 


cn > 


> 


l-i 


4-1 


> 


> 


l-l 


o 


C -H 


■H 





cn 


•H 


•H 


O 


•H 


M C 


c 


OO 


c 


c 


c 


00 


4-1 


-3 


3 


0) 


M 


p 


3 


CU 


ra 


01 




4-1 








4J 


2 


*-> .c 


X! 


re 


CJ 


* 


-C 


CO 




CO o 


U 


o 


•H 





o 


U 




> u 


u 




rH 


u 


u 






•H CO 


m 


i-i 


J3 


ra 


ra 


U 


0/ 


H <u 


cu 


0) 


3 


01 


01 


0) 


o 


0- CO 


ra 


J= 


fc 


cn 


CO 


^: 


1- 


01 


0) 


4-1 




0) 


Oj 


4-1 


3 


i-H OS 


PS 


o 


iH 


OS 


BS 


o 


O 



I 



u 

•H O 

•am 
ai 
B cn 

O C 

•H CO 

pa cu 

01 

T3 CN 
C I 
3 03 

fa 
W 

r^rJ 



5 



TO H 
l-i 

0) C 
T-( -H 
0) 

IH O 

o r. 
cn 
cn 

H CO 
CO 0) 

0> rH 

S 3 

4-1 

0) -H 
XI T3 
4-1 C 

cu 

r-l O. 

o X 

U-l [I] 

to J= 

l-< CJ 
CD l-i 

xi ra 

01 
CO 

cu 



B 

3 

z 

X 

0) 

c 



CO O 

u o 

M O iH 
•H 

n > u 

•H CO 

X CNI 

cn cu rs 

•a cc ON 

C I rH 

0) rH 

1-4 CO >4 



9 



H 

01 
J-l 

e 

3 
Z 



c 
o 

M-l -H 
O 4J 

3 

0) 4J 

0,i-l 

>N4J 

H cn 
C 



ON CN is m 



rs oo i— ) io 

O O rH ON 



rH iH 00 rs 

rH iH rH O 



o o o o 



o o o o 
o o o o 



O cn cn <■ 



O M0 O CN 
O ON CN CN 



CN -cr o in 



^rr on cn co 

OCACN -J 



O si CN O 



MO rH *CT CN 

rs rs o rH 



ntt cn cn on 

in CN iH rH 



<r m 10 o 



rs cn o rs 

O O rH CN 



CN MO rH cn 



cn ^r MO ON 

rH iH O rH 



O O O O 



o o o o 
o o o o 



O cn cn oo 



rH cn ON ON 

rH iH O ON 



m ^3- CN on 



cn <f cn o 

O ON rH ON 



rH IS ON H 



rs o cn \o 
rs rs on oo 



ON MO I — ^O 

00 cn cn cn 



CO 
















c 


(0 


tr. 




CO 


CO 


CO 







01 


01 




c 


01 


OJ 




•H 


•H 


-H 




o 


■n 


•H 




4-1 


4-1 


4-1 




•H 


4J 


4-J 




3 


■H 


•H 




4-1 


■H 


•H 




4J 


m 


co 


CO 


3 


CO 


CO 


cn 


■H 


u 


u 


0) 


4J 


u 


h 


01 


4-1 


01 


OJ 


•H 


•H 


01 


01 


■H 


to 


> 


> 


M 


4J 


> 


> 


V4 


C 


•H 


■H 


C 


CO 


•H 


■H 





M 


c 


c 


Ml 


c 


c 


c 


CO 




a 


3 


CU 


r^ 





to 


D 


01 






4-1 








4-1 


H 


XI 


J= 


ra 


o 


J3 


J3 


re 


ra 


CJ 


o 


CJ 


•H 


CJ 


o 


o 


> 


1-4 


1-1 




^H 


>-. 


>J 




■H 


CO 


CO 


u 


JO 


CO 


CO 


1-4 


H 


01 


01 


01 


3 


CU 


01 


01 


CU 


co 


co 


-C 


P-, 


CO 


en 


.c 




0) 


cu 


4-1 




cu 


OJ 


4J 


rH 


OS 


as 


c 


H 


sc 


P!i 


o 



w I 

4-1 



B-3 



-103- 



>N 






JS 






o 


C 




u 


•H 




CO 






01 


cn 




CO 


1-4 




a 


CD 




Zt 


i— 1 




H 







CD 


o 




O 






•H 


4-1 




DO 


c 




O 


a 




rH 


*-> 




o 


!fl 




•H 


P 




CQ 




o 




U 


1 




c 


1 




M-l 


c 




CO 


on) 


01 


•H 


*■•> 


u 


4-1 


O 


3 


3 


o 


1J 


4J 


1-1 


•H 


*n 




■a 


4J 


II 


B 


CO 




01 


e 


-J- 


a H 


■O 


X 




ON 


w 


o 


^H 


0) 


CO 


e 


■a 


01 




c 


•H 


* 


3 


l-J 


1-1 


fc 





o 




OO 


4J 


i>. 


0) 


CO 


rH 


^j 


■H 


H 


CO 


Uh 


03 


U 


0) 


U 




Q 


CD 


C 




ID 





a 


CD 


•H 


-3 


u- 


cn 

CO 


OS 


c 


•H 


93 


CO 


p 


H 


0) 


E 


Z 


X 



U 


— ' 


c 




to 


■H 


CD 


-o 




•H 


c 


(0 


00 


CO 


•a 


a> 


cn 


C 


c 


3 


0) 


i-j 


o 




co 


fS 



01 
XI 

e 

3 

z 



a 

o 

m -h 

O 4-1 

3 

0) 4-1 

a. -h 

>-, 4-> 

H co 
c 



cDonco 

vO fH ~3" 00 
On i— i NO CO 



>0 00 MN 

o> -J H m 

ON CN NO co 



<r on co m 

OCOHH 

r-~ <■ r-» co 



O cn co uo 

>J « H CN 
l/l CJ vO M 



CO N N O 

•a- oo ~? on 

*3 ON CN *;}■ 



00 CO CO o 

r*. on cm m 
r~ i-i cm <r 



■^ co m in 
i-l O f-~ o 

-3- CO O CO 



On -3- CM -3- 
iH r~ o r~ 
m vO ^H CM 



ON 


CN 


r^ 


m 


CM 


CO 


<r 


CO 


ON 


H 


CO 


r-» 


«3- 


nO 


-T 


rH 


m 


CO 


NO 


CM 


CO 


r^ 


CO 


CO 



in ro rs o> 
O u-i co t-H 

i-H .— I \D CN 



<r cn cn CT\ 



co cm m CM 

O on oo n 
iH CM o co 



ffi vO h- ^O 
00 CM CM CO 



cn 
















C 


cn 


cn 




CO 


cn 


cn 







CD 


u 




C 


0) 


cu 




■H 


•H 


■H 




o 


•H 


•H 




4-1 


4J 


4-1 




•H 


4J 


4-1 




3 


■H 


-H 




4-1 


•H 


•H 




4-1 


cn 


cn 


CO 


3 


cn 


cn 


cn 


•H 


l-i 


u 


cu 


4J 


S-t 


M 


01 


4J 


CD 


01 


■H 


•H 


0) 


cu 


•H 


cn 


> 


> 


M 


4-> 


> 


> 


l-i 


C 


■H 


•H 


o 


CO 


•H 


•H 


o 


i— i 


C 


c 


00 


c 


c 


c 


00 




3 


3 


cu 


tH 


3 


a 


0) 


0) 






4J 








4-1 


4-1 


43 


£ 


<0 


CJ 


Xi 


XI 


CO 


CO 





U 


o 


•H 


CJ 


CJ 


U 


> 


M 


hi 




i-l 


M 


i-i 




•H 


CO 


CO 


1J 


XI 


CO 


CO 


u 


1-1 


0) 


CD 


0) 


3 


CD 


01 


0) 


(X 


cn 


cn 


X! 


Cu 


cn 


cn 


Xi 




0) 


0) 


4-1 




0J 


0) 


4-1 


H 


a! 


os 


O 


■-I 


OS 


os 


o 



c 

3 
O 



cO 
C 
o 

•H 

4J 

z 



01 


1-1 

3 

o 



CO 




01 


u *-. 


3 O 


4-1 O 


•H i-l 


T3 


a ii 


01 


D. CM 


X r- 


w on 


iH 


T3 


CU >H 


•a i* 


c 


3 l-i 


tn O 


•4-1 


;>* 


^H CO 


•H C 


CO CO 


14 CU 


oi X 


T3 ^ 


01 


fn CO 
1 


u-l CQ 


O 


W 


cn ►-) 


CO < 


0) H 


S 


C 


01 -H 


X 


4-> C 


U 


O X 


UH C/} 


CO X 


l-i o 


cu 1-. 


X CO 


e oi 


3 CO 


Z 01 


OS 


X 


OI i-l 


T3 CO 


C O 


M -H 


00 


c o 


•H iH 


o 


CO 'H 


T3 PQ 


C 


01 l-i 


l-i o 


H M-4 


< 




CO 




ca 




W 




J 




M 




< 




H 





0) 

e 

3 

z 



c 
o 

14-1 -H 

4J 

3 

01 4-1 
O, -H 
>n 4J 
H CO 

C 



<T 00 CO CM 



in r^> o co 

tH i— I ON CN 



cn oo vx> m 



r-~ tH vo m 

HCN » O 



o o o o 



o o o o 
o o o o 



<r<f Hin 



ON c£) H 

on on co r^ 



on on <r cm 



co -a- on r^ 
ON ON oo oo 



on cnco 



in h oi on 

vO NO 00 NO 



<• CM CO ON 

in cm i— I i-l 



-J <t cO CO 



cm o in o 

O ON iH NO 



N ON CO NO 



uo no co r^ 

CM CM r-t <f 



o o o o 



o o o o 
o o o o 



<■ co in on 



r~ i-i CN on 

O iH O CO 



CN vO H N 



<r co m cm 

On 00 CM O 



O nJ rN H 



00 On i—l ON 

r^ no on O 



On no r^. no 

CO CN CN CO 



CO 
















c 


cn 


cn 




cn 


co 


CO 




o 


CD 


01 




c 


01 







■H 


•H 


•H 




o 


-H 


■H 




■U 


4-1 


4-1 




•H 


4J 


4-1 




3 


■H 


•H 




4-1 


■H 


•H 




4-1 


cn 


CO 


CD 


3 


CO 


cn 


(0 


•H 


u 


n 


CU 


4-1 


l-i 


1-4 


CD 


4-1 


01 


01 


•H 


•H 


01 


0) 


■H 


cn 


> 


> 


I-l 


4-1 


> 


> 


M 


c 


■H 


■H 


O 


CO 


•H 


•H 


o 


1-1 


c 


c 


00 


c 


C 


G 


00 




3 


p 


cu 


M 


3 


3 


CD 


01 






4-1 








u 


4-1 


X 


X 


CO 


CJ 


X 


X 


CO 


CO 


o 


CJ 


CJ 


•H 


u 


CJ 


u 


> 


1-1 


1-4 




i-l 


u 


M 




■- < 


CO 


CO 


U 


X 


CO 


CO 


iJ 


l-i 


0) 


01 


01 


3 


01 


0) 


CD 


0^ 


m 


cn 


X 


P-. 


01 


CO 


X 




01 


CU 


4-> 




01 


CD 


u 


.— 1 


OS 


OS 


o 


.— t 


as 


OS 


o 



B-4 



-104- 



X c 

•H 

x; 

CJ CO 

u u 

CO to 

QJ r-t 

CO r- i 

01 o 



co c 
o CO 
•h vj 



id 



c 

o 

U 
U I 

O I 

m-i en 
C 
en oto 

3 



4.1 O 

3 O 



■a u ii 
c en 

u c <• 

a. i— i vc 

x a, 

HUH 

O 

■O Sh 

en en tu 
-a i> 

C -H « 

3 M H 

tv o c 

00 VI 

s, 1) CO 

rH VI rH 

iH CO CM 

CO CJ 1) 

I- Q 
0) c 

•O OQ 

0) *H u3 



CM 

i-t a 



oi is 

a o v 
o 

C CO 

HUT, 

•H C 

cn M co 
t3 oi en 

C C 3 
01 M o 



H CJ 



st 

I 
ca 

w 
5 



3£ 



CO 
0) 



CO 



sr 

cr. 



u 

0) 
XI 



c 
o 

HH -H 

O VI 

3 

1) v> 

D. -H 

>s vi 

H en 

C 



ro i-l 


CO 


CI 


fS 


CN 


ST 


sr 


st f-l 


H 


rH 


rH 


CN 


CM 


r-l 


sr <n 


o 


CN 


en 


en 


rs 


en 
















CN u-| 


H 




rH 


CO 






</> 














oo m 


CN 


o 


CM 


rH 


CO 


in 


sr in 


rs 


lO 


o> 


sr 


rH 


m 


\D m 


o 


cn 


o 


rs 


X) 


CN 
















cn m 


rH 




rH 


CN 






</> 














in rs 


CTi 


cn 


rH 


rH 


CO 


C^ 


O vO 


<r 


CN 


CN 


m 


O 


OS 


in ro 


cc 


cn 


o 


cn 


rs 


CM 
















cm m 






rH 


CM 






</> 














r~i in 


rs 


O 


CO 


in 


00 


m 


ST CM 


00 


vD 


en 


O 


rs 


rs 


vd cn 


CM 


<r 


rH 


rs 


rs 


CM 
















cn in 


rH 




rH 


esi 






</> 














i-t oo 


CN 


rs 


CN 


m 


st 


O 


m cn 


Is 


00 


ST 


CO 


lO 


en 


rs vo 


rH 


ST 


rH 


<t 


o 


CN 
















cn m 


H 




rH 


CN 


rH 




</> 














-a- oo 


m 


ST 


en 


ro 


en 


CO 


<y\ 00 


rs 


CI 


m 


es 


cr* 


CM 


rH CN 


o 


m 


eji 


rH 


rs 


CN 



<f CN CO a\ 

m CM H H 



cjn vo r*» \o 

oo n m ro 



en 
















C 


en 


m 




en 


en 


en 




o 


0) 


0) 




C 


0) 


01 




•H 


■H 


•H 







•H 


•H 




V> 


u 


4J 




•H 


4H 


l-l 




3 


•H 


•H 




v. 


•H 


•H 




4J 


en 


en 


en 


3 


en 


CO 


eo 


•H 


M 


(J 


0) 


!H 


H 


rJ 


0) 


u 


0) 


0) 


•H 


■H 


1) 


0) 


•H 


CO 


> 


> 


IH 


4J 


> 


> 


rJ 


c 


•H 


•H 


o 


en 


•H 


•H 





rH 


c 


C 


00 


C 


C 


C 


00 




p 


P 


0) 


H 


o 


P 


11 


0) 






u 








■u 


Vt 


X 


X 


CO 


o 


x; 


J3 


CO 


ft) 








o 


■iH 


o 


u 


U 


> 


H 


rJ 




rH 


u 


1-1 




■H 


CO 


CO 


M 


XI 


m 


CO 


IH 


H 


01 


01 


01 


3 


<u 


0/ 


'11 


CH 


to 


en 


J3 


co- 


en 


CO 


x: 




a 


01 


vi 




01 


CU 


4J 


rH 


Oi 


p3 


O 


rn 


pej 


Kl 


O 



CO 
-O 

3 

O 



cj 
C/3 



CO 

rS 



0) 

o 

IH 

3 

o 

CO 



CO 




CJ 




iH 




3 




4J 




■H 


r-** 


T3 


o 


C 


o 


'II 


rH 


cx 




X 


II 


ej 






CN 


■a 


r-* 


CD 


CT, 


"0 


rH 


c 




3 


r- 


It, 


tt 


>, 


u 


rH 





rH 


it-i 


cr 




M 


en 


<V 


C 


T) 


tei 


tu 


01 


Li- 


S3 


'+H 







<r 


en 


DO 


c 




cr 


III 


0) 


rJ 


>: 


pn 




< 


01 


H 


X 




4H 


a 




■H 


IJ 




o 


C 


4H 


^ 







i/; 


43 


M 


CO 


'.V 




rQ 


X 


B 


CJ 


3 


n 


Z 


CO 




01 


X 


en 


01 


CJ 


*n 


BS 


C 





cn en 
c 



0) 

n 

H 



1-, 
0) 

■s 

3 
S3 



C 

o 

llH ^H 

4J 

3 

01 u 

a. -h 

r>< 4J 

H en 
c 



m rH oo icn 



r-* rs o\ m 

CT, CJ\ rH i£> 



r^ in cn m 



m cn vc rH 

O O CN rH 



o o o o 



o o o o 
o O o o 



st cn m m 



in O^ rH CN 

ooiin>i 



CO rH O rs 



o m oo o 
O o cn m 



^ Oi lO c^l 



r-~ cr> yo m 
oo r- cn mo 



st cn cn CT\ 

in CN rH rH 



o cn cti o 

<-n rH cn m 

cn sr o o 

rH rH rH rH 



ct\ \c <~r, st 



HO vO f^ lO 
O rH 00 00 



o o o o 



O O O O 
O O O O 



sr O r^ CN 



rH in O CN 
rH r-i ^ C7* 



CO is CO H 



rH in rH is 

rH o m is 



m cn co sr 



cn o cn yco 

O OS rH I — 



c ^o rs \£> 
00 CN cn cn 



c 


eo 


CO 




en 


en 


en 




o 


01 


CJ 




c 


0) 


OJ 




•H 


•H 


■H 




o 


•H 


■H 




4J 


+J 


«J 




•H 


VI 


VI 




3 


•H 


■H 




4J 


•H 


■H 




iH 


eo 


CO 


en 


3 


en 


in 


CO 


■H 


rJ 


hi 


<u 


4-> 


M 


>j 


0) 


4-» 


0) 


cu 


■H 


TH 


01 


01 


•H 


CO 


> 


> 


u 


u 


> 


> 


rJ 


c 


•H 


■H 


O 


en 


■H 


tH 





rH 


C 


a 


00 


C 


C 


c 


00 




3 


p 


CU 


H 


^ 


3 


CJ 


01 






4J 








VJ 


u 


X 


X 


CO 


o 


X 


X 


CO 


CO 


O 


a 


CJ 


•H 


o 


CJ 


u 


> 


M 


u 




rH 


rJ 


r4 




tH 


CO 


CO 


M 


Xl 


CO 


CO 


u 


IH 


01 


01 


0) 


3 


0) 


1) 


0) 


CH 


CO 


cn 


rC 


Ph 


in 


in 


X 




0) 


ii 


4H 




tu 


01 


V 


rH 


Pi 


pS 


o 


rH 


ev- 


p; 


o 



B-5 



-105- 



TABLE B-5 . Trends in Mean Federally Funded Expenditures for 

Research in Life Sciences not Elsewhere Classified, 
by Carnegie Commission Categories of Institutions — 
Constant Dollars in Thousands (NIH E&D Deflator, 
FY 1964 = 100) 



Type of 
Institution 



Number 



Fiscal Year 



1968 



1970 



1972 



1973 



1974 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



54 


$157 


$145 


$121 


$203 


$291 


22 


106 


218 


234 


325 


575 


13 


161 


104 


96 


287 


212 


19 


214 


121 


7 


5 


15 


89 


181 


250 


177 


112 


63 


26 


479 


591 


507 


178 


89 


27 


113 


175 


79 


136 


111 


36 


21 


65 


17 


49 


10 



Source: National Science Foundation. 



TABLE B-5A . Trends in Index Numbers for the Means of Expenditures 
for Research Life Sciences not Elsewhere Classified 
Shown in TABLE B-5 (Means for FY 1972 = 100) 



Type of 
Institution 



Number 



1968 



Fiscal Year 



1970 



1972 



1973 



1974 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



54 


130.1 


129.3 


100.0 


167.9 


240.4 


22 


45.2 


93.2 


100.0 


138.9 


245.8 


13 


168.1 


108.5 


100.0 


299.3 


221.6 


19 


3014.7 


1698.8 


100.0 


65.4 


208.3 


89 


102.1 


141.4 


100.0 


63.5 


35.8 


26 


94.5 


116.8 


100.0 


35.2 


17.5 


27 


141.9 


220.4 


100.0 


171.2 


140.1 


36 


125.2 


394.6 


100.0 


294.5 


63.2 



B-6 



-106- 



>> 






oo 






c 






rH 









in 




x 


U 




o 


B 




>,rH 




oi 


rH 




c- 


o 
Q 




C 






•H 


u 
C 




X 


ffl 




o 


vj 




u 


in 




ro 


c 




III 







en 


u 




01 


1 




os 


1 

m 




u 


C 







ora 


U-i 


•H 


r — * 




■U 


o 


«: 


3 


o 


0) 


-U 


rH 


H 


•H 




3 


U 


II 


4-1 


n 




-H 


C 


<r 


"0 


M 


uO 


c 




con 


01 


UH 


rH 


c- 







X 

UO 


in 

01 


E 


-a 


■H 


• 


01 


r-l 


P 


•a 


O 





c 


00 


4J 


3 


Ol 


to 


o- 


4-1 


^H 




CO 


14-4 


Pn o 


01 


rH 




a 


.-1 


c 




ro 





a 


i- 


■rl 


"3 


0) 


in 


OS 


-a 


to 




0) 


■H 


X 


fc. 


1 


M 

Z 


c 





— ' 


ro 


o 




01 




in 


s 


01 


-o 




■H 


c 


c 


to 


n 


i-i 


0) 


CO 




c 


3 


10 


1-1 


O 


■a 

c 


IB 
U 


P 


11 






H 


^ 


c 


H 


XI 


■H 



<T 


oo 


O CM vO 


U3 


rH 


O "A 


r» 


r^ 


inoo-f 


rH 


kD 


00 vO 


ON 


i-l 


fi 


CM 


-J 


i-l 



1-4 

01 

x 

e 

3 

z 



c 
o 

14-4 -H 

4J 

3 

01 4-1 

;>> 4J 

E-i 01 

c 



00 -3- ON ON 
ON ON ON f* 



r~ en oo cm 
cm oo 00 r~ 

CM ~d- rH 



ON 00 ON *4* 
i-l PI 


i-l vO 
CM vO 
CM <3r 


CO 

rH 


</> 







ro <j c*i n 
00 in iH en 

HflH 


0\ CM 

cn en 

CM 1/N 


--1 


u-i 


<J> 









-T 


MO 


rH 


CM 


^H 


rH 


r^ 


--T 


O 


in 


CO 


<T 


N* 


en 


ON 


SD 


cm 


en 


rH 




r^4 


in 


rH 




</> 

















<T -41 i-l 
vO CM i-l 

HP1H 



»» IN CO CN 
KN CM H H 



ON iH 00 M3 

IN N H n 



On vO r* mo 
00 cm cm en 



in 
















c 


in 


in 




in 


in 


in 







Ol 


0) 




c 


0J 


01 




■H 


•H 


■H 







■H 


-H 




4-1 


4-> 


4-1 




•H 


4-1 


4-1 




3 


•H 


•H 




4-4 


■H 


•H 




4-1 


in 


in 


CO 


3 


01 


in 


in 


•H 


Si 


i- 


01 


4-1 


rJ 


M 


04 


4-1 


01 


0) 


■1-t 


■H 


0) 


01 


•H 


n 


> 


> 


l-l 


4-1 


> 


> 


M 


c 


•H 


•H 


o 


in 


■H 


■H 


o 


H 


c 


c 


60 


c 


C 


C 


60 




3 


3 


0) 


M 


D 


S 


01 


04 






4-1 








4-1 


4-1 


X 


Xi 


H3 


CJ 


x 


X 


ro 


CO 


o 


U 


o 


•H 


j 


O 


O 


> 


M 


M 




r-l 


u 


M 




■H 


ro 


CO 


r-l 


Xl 


M 


ra 


M 


u 


01 


01 


01 


3 


01 


0) 


01 


Cu 


n 


in 


J= 


P* 


in 


in 


X 




0) 


0) 


4-1 




0) 


01 


4-1 


rH 


« 


OS 


o 


iH 


OS 


OS 


o 



c 

3 


t- 

01 

o 

c 



ro 
Z 



V4 











/ ^ 


in 


o 


0i 


o 


u 


rH 


3 




4-1 


II 


•H 




TO 


^4 


C 


r~. 


01 


ON 


O. rH 


X 




W 


!h 




U. 


t3 




OJ 


M 


-a 





c 


'-H 


3 




In 


if: 




C 


>. 


ro 


rH 


01 


— I 


S 


CO 


• — 


u 




01 


x> 


— 1 


I 


a. 


211 


1* 






W 


U-I 


-J 


c 


S 


in 


H 


c 




rO 


C 


CI 


■H 


E 






ro 


u 


3 





o 


IH 


X 




03 


in 




M 


^ 


01 


tA 


.11 





h 


rH 


3 





Z 


J= 




'J 


X 


r»N 


01 


in 


■a 


Pn 


c 




H 


c 




•H 


c 




■H 


-= 




u 


in 


H 


■a 


ro 


ro 


"I 


01 


X, 


u 


01 



\0 fO CM H 



ma^OH 



CM H CO n 
0** <J\ 00 •—) 



xO CO I — *nD 



r*. oo oo O 
o\ o^> o^ o** 



a. co rs o 



i) 

■i 

3 
Z 



<i 




O J-i 


1 




3 
01 4-1 


30 




D. -H 
P^ 4-1 


I--' 




H in 


-I 




C 


CC 




M 


< 






— 







CM CM vO ~3f 
O O O ON 



o o o o 



CM en CM ON 
O O O ON 



o o o o 



o o o o 
o o o o 



O -3* vO vO 



o o o o 
o o o o 



CM CM O ^T 



in cm i— i i— < 

ON ON CM 00 



^O ON ON O 



oo -a- us m 

O rH ON O 



O ON sC i— I 



in cm in -4j- 

O ON ON O 



CM VJO cm m 



on en i on 

O rH O 00 



in rH cm en 



m -Of O r~ 
oo oo cm en 



-j in n oi 

4n CM rH rH 



00 00 -o- ON 
in m \C <r 



ON uO r-~ u3 
oo in cm en 





rH 


h- 






^ 


M 




n 
















c 


in 


in 




in 


01 


140 







0) 


01 




B 


0) 


0) 




•H 


■H 


•H 




O 


•H 


■rl 




4-1 


4-1 


4J 




•rl 


4-1 


4J 




3 


•H 


■H 




4-4 


■H 


•H 




4-1 


to 


in 


CO 


3 


01 


01 


a 


•H 


u 


H 


01 


4-1 


r-l 


rJ 


01 


4-J 


01 


0i 


■H 


•H 


01 


OJ 


•H 


to 


> 


> 


r^ 


4-1 


> 


> 


U 


c 


•H 


•H 


c 


01 


■H 


■H 





rH 


3 


C 


60 


c 


c 


c 


oc 




5 


30 


01 


r-l 


30 


p 


01 


0) 






4-1 








4-1 


4J 


J3 


X 


CO 


u 


X 


X 


CO 


CO 


U 


o 


o 


•H 


o 


u 


CJ 


> 


r-l 


u 




rH 


u 


u 




■H 


eo 


10 


r-l 


Xl 


ro 


ro 


1-1 


M 


01 


01 


0) 


3 


01 


01 


ll 


0. 


in 


If, 


rC 


o, 


0] 


01 


X 




V 


01 


4J 




o 


01 


Hi 


H 


OS 


OS 


o 


rH 


OS 


OS 


O 



B-7 



-107- 

APPENDIX C 



ENROLLMENT TABLES FOR INSTITUTIONS CLASSIFIED BY CATEGORIES 
OF THE CARNEGIE COMMISSION ON HIGHER EDUCATION 



TABLE C-l. Trends in Mean Degree-Credit Enrollment by Carnegie Commission 
Categories of Institutions 

TABLE C-1A. Trends in Index Numbers for the Means of Degree-Credit Enroll- 
ment Shown in TABLE C-l (Means for FY 1972 = 100) 

TABLE C-2. Trends in Mean Enrollment for Advanced Degrees in All Fields by 
Carnegie Commission Categories of Institutions 

TABLE C-2A. Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in All Fields by Carnegie Commission Categories Shown 
in TABLE C-2 (Means for FY 1972 = 100) 

TABLE C-3. Trends in Mean Enrollment for Advanced Degrees in Biomedical- 
Behavioral Sciences by Carnegie Commission Categories of 
Institutions 

TABLE C-3A. Trends in Index Numbers for Means of Enrollment for Advanced 
Degrees in Biomedical-Behavioral Sciences Shown in TABLE C-3 
(Means for FY 1972 = 100) 

TABLE C-4. Trends in Mean Enrollment for Advanced Degrees in Biological 
Sciences by Carnegie Commission Categories of Institutions 

TABLE C-4A. Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Biological Sciences Shown in TABLE C-4 (Means for 
FY 1972 = 100) 

TABLE C-5. Trends in Mean Enrollment for Advanced Degrees in Health Profes- 
sions by Carnegie Commission Categories of Institutions 

TABLE C-5A. Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Health Professions Shown in TABLE C-5 (Means for FY 
1972 = 100) 

TABLE C-6. Trends in Mean Enrollment for Advanced Degrees in Psychology by 
Carnegie Commission Categories of Institutions 

TABLE C-6A. Trends in Index Numbers for the Means of Enrollment for Advanced 

Degrees in Psychology Shown in TABLE C-6 (Means for FY 1972 = 100) 



C-j 



-108- 











to oo co >o 


c- oo to to 




















•* 


mnit iO 


GO CJ) O 00 








t 


>como 


IO Cn C- rH 






r~ 


onosi 


«T O cn t— 








p^ 










o> 












en 


O rH rH Cn 


IO tO CM tO 






rH 


OHcno 

i-H rH l-H 


OOIOI* 
CM CM rH rH 

O *J- CO O 




rH 
1 

u 

0J 
-J 




i-i 


o o o en 

l-H rH rH 


o o o o 

i-H i-H i-H rH 






tO 


\onoiN 


CM 00 vO M3 




ca 




to 


C-. O CO t> 


lo cm cn o 






r- 


vO *» CO en 


tO rH LO LO 




< 




e'- 


• • • • 


■ • • I 






o> 








H 




en 


o cm en en 


N-JOH 






rH 


O i-H 00 O 

rH rH rH 

l-H LO CM CO 


OOlOl^ 
CM CM rH rH 

rH IO rH CO 




C 
•H 




r-l 


o o en en 

f-H rH 


o o o o 

HHHH 


a 




CM 


OlOHO 


IO rH CO tO 




O 




CM 


o o o o 


o o o o 


o 




c- 


LO CM CT1 O 


CO o to to 




J= 




t-» 






•H 




cn 








CO 




en 


o o o o 


o o o o 


W 




rH 


OHOOH 


CT> CO Oa *3" 








rH 


o o o o 


o o o o 


in 


ri 




I-H rH rH 


HNrlH 




«•> 


rl 




rH rH rH rH 


rlHHH 


•H 


«) 










e 


CO 








I 


<D 










4> 


4> 








>< 










s 


>H 








o 






r~ vo r- o 


00 "". o> "" 




rH 










u 


' rH 


rH 


cn vo to c- 


Orlr-Ol 




i— I 


rH 


rH 


CO tO O VO 


\D LO «3" *© 




CO 


r— 


CTlNr-O 


C~ 1*- co Cn 




o 


CO 


c~ 






ID 


u 


en 








rl 


U 


en 


to en co o 


OlNCtv 


•H 


10 


rH 


O CM CO i-H 


en oo oo to 




C 


in 


t-< 


o o en o 


cn o cn cn 


00 


•H 




rH rH rH 


rH CM rH rH 




w 


■H 




rH rH i-H 


rH 


4> 


U. 












Ph 








c 












+J 










u 


IH 










•H 


H-l 








rt 


O 




CT> VO CM VO 


rH CJ> CO IO 




■a 


o 








CJ 




o 


rH Cn M3 t>- 


CM LO CO H 




4) 




o 


tO CM LO CO 


cn o cn oo 




e 


r-~ 


vO 00 O CO 


•5* © IO o 




r< 


e 


r- 






X 


s- 


O) 


•*•*•>•• 


m m at « 




U 


rl 


en 


O vO O 00 


CM to 1* o 


XI 


<D 


i-H 


O rH 00 O 


CO o co to 




1 


4) 


r-< 


o o en en 


cn cn cn cn 




H 




rH rH rH 


rH CM rH rH 




4) 


H 




rH r-l 




*J 












4) 










c 


rH 










rl 


rH 








<u 


rH 










00 


rH 








£ 


cd 




nsdo 


O rH 00 vO 




4) 


CO 








rH 


tl< 


C71 


to oo >o r- 


Niomo 




Q 


u. 


en 


13* CM CM CM 


to r— oo o 


i-H 




^ 


to r- »© to 


to oo to en 








vO 


• • • • 


• a • • 


O 




Cn 








HH 




en 


OlWr-"t 


c- oo cn to 


r. 




rH 


O rH 00 O 


t-. "* t-~ rH 




O 




r-l 


en o en en 


CO CO CO 00 


G 






rH rH rH 


rH CM rH rH 










l-H 




Uj 












in 

a 










*-> 












CO 










•HCO 






O^tHlfl 


CM O rH CM 


• 


4) 










"0 Ifl 




o 


«3* ro *3* c- 


vO ^ tO CM 


tn 


s 




vO 


Cn tO CM rH 


•a- r-. cn oo 


0). c 




vO 


OONO<t 


HlOOIC- 


o 






M3 


• • • 


• • • • 


u o 




Cn 






•rl 


4> 




en 


NOOO 


rH tO rH C~ 


U -H 




r-H 


cn i-h co en 


■* o to a> 


+J 


-C 




rH 


en o en co 


C- C- c- \o 


1 +J 






rH 


rH CM rH 


in 


+J 






rH 




4> 3 










•H 












0) +J 










+J 


rl 










l-t -H 










rt 


O 










00 4-> 






rH CM CM tO 


rH rH CM O 


r> 


4H r^ 










4> W 




l/» 


CM 00 tO O 


irnorv n 


CO 


o 




LO 


o o co o 


WOoiN 


o g 




>o 


toic-co 


co vO tn ai 




in o 




M3 


• • • ■ 


■ * a • 


1— t 




Cft 


M «* «a at 


*«•>•( 


c 


f-l rH 




en 


en co vo o 


•a- o •!)■ m 


c 




rH 


aot-oo 


CM 00 CM CO 


o 


4) 




rH 


co en co co 


O vOOO 


CO 4-1 






rH 


rH rH rH 


•H 


-O II 










0) o 










+J 


B 










£ 








C8 


3 CM 








tn 








O 


Z r- 








C <D 


f-< 






3 


en 


rl 






•H -H 


03 


LO CM tO © 


CM C~ t^ CO 


13 


X rH 


4) 


LO CM tO O 


CM C~ [^ CO 


rl 


43 


lo cm t-i c< 


en CM CM to 


UJ 


4) 


r3 


LO CM rH CM 


cn CM CM to 


V) O 


£ 








T> >^ 


e 






T> 60 


3 






rl 


c a. 


3 






a <o 


z 






O 


t— i 


z 






4> 4-> 








>4H 


rl 








(-• ra 




rH 


rH 




e o 




r-l 


r-l 


t- U 




rH rH 

in 


rH t-i 


rl 
4> 


•H 1H 




r-l r-l 

m 


rH rH 






C in in 


m m m 


+J 


in tn 




G m in 


in m in 


• 




O <L> 4> 


COO) 


C 


T3 g 




O 4) 4) 


G 4) 4) 


rH 






•H -H aft 


O -H -H 


4) 


S co 




•H -H -h 


O -H .H 


i 






4-> 4-> 4-> 


•H(JU 


U 


4> 4) 




«-> *-> *-> 


•ri *i +J 


u 






3 -H -H 


*J -rH .H 




rlS 




3 'H -H 


+J -H tH 








*j m in m 


3 w <n tn 


rH 


H '-' 




+J tn in tn 


3 tn tn in 


w 




c 


■ H !H rl 1) 


*-> r. r. 4) 


CO 




c 


•H SH rl 4) 


•(-> rl rl 4) 


J 




o 


+j 4> a, .h 


•rl 4) 4) -H 


C 




o 


<-> 4) 4) -H 


•H 4) 4) tH 


§ 




LH -rl 


in > > r. 


■P > > rl 


o 




4H -H 


in > > In 


+-> > > rl 




o +-> 


C -H -H O 


tn -H -rt O 


•H 


• 




O 4J 


G -H -H O 


m 'H -rH o 


H 




3 


rH C G 00 


C C C OO 


*J 


< 




3 


i-H C C 00 


G G G M 




m +j 


3 D 4> 


t-H 3 S 4) 


CO 


rH 




4) 4-> 


D S 4) 


rH 3 3 4) 




a-H 


4> *j 


•M 


z 


1 




CX-H 


4> 4-> 


+J 




>>u 


l-> JC -C CO 


u j3 js a) 




u 




>>« 


*-> J3X CO 


O JS JS CO 




H in 


« o o o 


•H u u u 








H « 


tO U O U 


•H O U CJ 




G 


> rl rl 


rH rl rl 


■ • 


m 




C 


> rl rl 


rH )h rl 




HH 


•rl CO CO ri 


^D tO CO ri 


4) 


rJ 




t— i 


■H CO CO In 


^ CO CO ri 






nl)III 


3 4) 4) 4) 


O 


PQ 






ri 4) 4) 4) 


3 4) 4) 4) 






a, « in j2 


o, in tn js 


rl 


< 






Bu in in j3 


cl, m in -G 






4> 4) <-> 


4) 4> r> 


3 


H 






4) 4) +J 


4) 4) +J 






rH Q. OS O 


rH Oi Pi o 


O 






rH Qd « O 


rH QC 0, O 






rH 


i-H 


OT 






r-t 


l-H 










< 


< 


CO 










< 


< 



C-2 



-109- 



x 

t/i 
i3 

rH 
41 
•H 



to 


in 


VI 


4) 


c 


1) 


o 


r. 


■H 


oo +J 


0) 


3 


P 


4J 




■H 


T5 


4-J 


0) 


l/l 


U 


C 


a 


i — i 


TO 




> HH 


-o 


o 


< 






t/1 


(-1 


o 


o 


•H 


4h 


h 




o 


t-i 


oo 


c 


o 


0) 


*J 


E 


TO 


rH 


U 


^H 




o 


c 


r< 


o 


c 


■H 


w 


tn 




t/i 


B 


■H 


ra 


6 


u 


e 


S 


o 




u 



-a 
o 



H U 



gj 



3 



c 

o 

m -h 

o *-> 

3 

>>^ 



HCOOON 
O LO rH CO 
CO CO rH O 



00 O N CO 
C"0 «3" Ol CO 



rg ro rg rg to lo rg rH 



rg *Lf CO rH 
E-- rH -^ tO 

NOOHO 

IN t^N N 



lo rg lo eg 

a> ^o t^ n 

r- CO rH o 



LO LO «* Ol 

\o ai vo n 
rg ^ en co 

to lo rg rH 



^* t-^ <o ■o 

NOI ro H 
og rt rH r- 



(S KIN N 



OHN N 

r- ^t o ai 

t-» O rH CO 



cn r^- o o 
oi t-^ \o o 
co ro ai lo 



rg to rg rH 



i—t ai oi r- 
o o rg o 
r*- cr> o co 



(O LO rg rH 



lo to r-. eg 



m rON H rg lo rg i— i 



r-»- -st- o rg 
co to o r>. 
lo rx- CTi r~- 



rg lo r^ rH 

^H NN 

r- ai ^j- to 



rg tO rH rH 



rg Ty rg ph 















TO 


00 ■"* 


CO 


rO 


ro 


ai ai 


O) 


+J 


ro co 


ro 


O 


!-H 


LO LO 


00 


CO 


rt lo 


r^ 


r^ 


«* 


LO o 


o 


C 


co ro 


rH 


i— I 


ro 


tj- co 


rH 


O 

•H 

♦-> 

TO 
U 

D 


o o 


CO 


00 


i— 1 


rH CO 


00 


t3 


LO (V) 


f-H 


i— i 


f^ 


CO co 


r-j 


PJ 

H 

o 

MH 


t— I 


l — 1 






HH 1— t 




|H 


l/l 












41 


C tn 


U1 




tn 


l/l </l 




■!-) 


O 4) 


01 




C 


41 4) 




c 


•H -H 


•rH 




o 


•H -H 




41 


4-> +J 


4-> 




•H 


4-> 4-> 




CJ 


3 -H 


'H 




+J 


■H -H 






4-J l/l 


tn 


tn 


3 


tn t/i 


tn 


rH 


■H Sh 


Sh 


ai 


4-> 


Sh Sh 


0) 


TO 


4-> 0) 


0) 


■H 


•H 


4) 4) 


•H 


c 


tn > 


> 


t-H 


4-> 


> > 


rH 


o 


C -H 


■H 


o 


l/l 


•H -H 


O 


• H 


>-i c 


c 


oo 


c 


c c 


OO 


■I-J 


3 3 


1) 


(— 1 


3 3 


41 


TO 


0) 




4-> 






■rJ 


z 


+-> X X. 


TO 


U X 


TO 




TO U 


u u 


•iH 


u o u 




> r< 


u 




rH 


!h Sh 






■H TO 


TO 


U 


XI 


TO TO 


tH 


41 


Sh 4) 


a) 


41 


s 


4) 4) 


41 


O 


a. in 


«£ 


<X 


in l/l X 


rH 


0) 


o 


■M 




4) 41 


+-" 


3 


HKtfO 


rH 


Pi OS o 


o 


f— 1 

< 






f-H 

< 






en 

TO 



o 
o 



< I~- 



41 
4) r- 

rH O 

OOMH 

41 
Q tn 

c 

T3 TO 
41 41 
L> S 

C "— ' 
TO 
> ro 

■a i 

< u 

!h W 

o ^ 

< 

+J t-, 

c 

41 C 
S-H 
rH 

r-l C 

O 2 

tH O 

c x; 

BJ CO 

m tn 

O 41 

•H 

tn rH 

C O 

TO 00 
41 41 

TO 
41 t_) 
XI 

4-1 C 

o 

rH -H 

in 
|-H tn 

•H 

m S 

H e 

41 O 
XI u 

6 

3 41 

2 -h 

oo 

K 41 

01 C 
T3 tn 

C TO 
h u 

C >, 

•H X 
l/l l/l 

C rH 

41 41 

tH .H 

H (jl, 



rH 
41 

X 

e 

3 
Z 



o 

4h -h 

O -M 

3 

41 4-1 

CVh 

H i/i 
C 



O rH vO LO 



LO LO tO rH 



rH rH OO CO 
O O CT. O 



o o o o 



O CTI O IO 

o cn o o 



o o o o 



o o o o 
o o o o 



CO N (N SO 



o o o o 
o o o o 



ro o co oo 



OHHOt 

o o o cn 



HKlvOH 



O O lo to 
o o o ai 



CTi CTt CTl O 



O ro o ro 

o o o ai 



<3- LO LO O 



*3- r^ ai co 

ai ai cn oo 



rH r— cn r~- 



C^ CNl rrj- CTl 

OlOOlCO 



ro CTi lo rg 



oo tNi r^. r-~ 

oo CTt oo r^ 



o lo o co 



ro r-~ oo r-- 

ai cn oo oo 



cn ■* r- oo 



«d- cn ro co 

00 OO 00 1^- 



oiom n 



r^ co o ro 
co ai oo oo 



O CD CO OO 
LO O-l rH rH 



rO ro CTi CTi 



rH rH r<I CO 
t-^ CO CO CO 



to 




C tn in 


in tn tn 


O 41 41 


C 41 41 


•H -H -H 


O -H -H 


■M +-i •»-> 


•H *-" +J 


3 -H >H 


+J -H 'H 


+h t/i tn tn 


3 i/i tn t/i 


•H (H tn 41 


■P Sh tn 41 


+-1 41 41 -H 


• H 41 41 -H 


m > > rH 


4-1 > > r. 


C -H -H O 


l/l >H 'H O 


rH C C 00 


c c c oo 


3 3 41 


h33J) 


41 •*-> 


4-> 


■M X X TO 


U X X TO 


TO U O U 


•H O U U 


> tH rH 


rH In tn 


•H TO TO rH 


X TO TO tn 


tn 41 41 41 


3 41 41 41 


a, tn l/l x 


D, tn tn x 


41 41 +-" 


41 41 4-> 


HKOlO 


rH OS OS O 



C-3 



-110- 



TABLE C-3 . Trends in Mean Enrollment for Advanced Degrees in Biomedical- 
Behavioral Sc 
Institutions 



Behavioral Sciences by Carnegie Commission Categories of 



1967 


1968 


1969 


1970 


1971 


1972 


1973 


221 


231 


234 


240 


245 


267 


291 


305 


316 


320 


330 


332 


371 


393 


225 


231 


243 


246 


254 


263 


293 


130 


141 


137 


141 


146 


160 


180 


303 


328 


353 


372 


382 


419 


434 


594 


632 


678 


719 


737 


796 


828 


256 


273 


282 


300 


308 


340 


357 


119 


140 


163 


164 


171 


195 


196 



Type of Number Fall Term of Fiscal Year 

Institution 



All Private Institutions 49 

Research Universities I 19 

Research Universities II 12 

Other Categories 18 

All Public Institutions 70 

Research Universities I 21 

Research Universities II 21 

Other Categories 28 



Source: National Center for Education Statistics. 



TABLE C-3A . Trends in Index Numbers for Means of Enrollment for Advanced 
Degrees in Biomedical-Behavioral Sciences Shown in TABLE C-3 
(Means for FY 1972 = 100) 



Type of Number Fall Term of Fiscal Year 

Institution 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 





1967 


1968 


1969 


1970 


1971 


1972 


1973 


49 


82.8 


86.4 


87.6 


89.9 


91.6 


100.0 


108.7 


19 


82.2 


85.1 


86.3 


89.0 


89.6 


100.0 


106.0 


12 


85.4 


87.9 


92.3 


93.4 


96.3 


100.0 


111.4 


18 


81.4 


87.9 


85.8 


88.2 


91.2 


100.0 


112.5 


70 


72.2 


78.2 


84.3 


88.7 


91.2 


100.0 


103.6 


21 


74.6 


79.4 


85.2 


90.4 


92.6 


100.0 


104.1 


21 


75.3 


80.3 


82.9 


88.2 


90.5 


100.0 


104.8 


28 


60.9 


71.7 


83.5 


84.3 


87.9 


100.0 


100.6 



C-4 



-Ill- 



49 


117 


123 


120 


122 


118 


119 


125 


19 


184 


193 


196 


200 


191 


189 


199 


12 


80 


87 


77 


75 


73 


71 


77 


18 


72 


73 


68 


72 


71 


75 


80 


70 


179 


190 


199 


206 


212 


215 


215 


21 


334 


356 


361 


376 


381 


392 


400 


21 


159 


161 


181 


185 


186 


187 


184 


28 


78 


89 


93 


97 


107 


104 


100 



TABLE C-4 . Trends in Mean Enrollment for Advanced Degrees in Biological 
Sciences by Carnegie Commission Categories of Institutions 3 

Type of Number Fall Term of Fiscal Year 

Institution 1967 1968 1969 1970 1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



Source: National Center for Education Statistics. 



TABLE C-4A . Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Biological Sciences Shown in TABLE C-4 (Means for 
FY 1972 = 100) 



Type of Number Fall Term of Fiscal Year 

Institution 1967 1968 1969 1970 1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



49 


99.0 


103.5 


101.3 


103.3 


99.3 


100.0 


105.8 


19 


97.1 


101.7 


103.7 


105.4 


100.6 


100.0 


105.2 


12 


112.3 


121.9 


107.9 


105.4 


102.1 


100.0 


107.6 


18 


95.8 


96.8 


90.7 


96.2 


94.4 


100.0 


106.2 


71 


83.3 


88.5 


92.9 


96.2 


98.9 


100.0 


100.1 


21 


85.4 


90.8 


92.2 


96.0 


97.2 


100.0 


102.2 


22 


85.3 


86.0 


97.2 


98.8 


99.8 


100.0 


98.8 


28 


74.5 


85.3 


88.8 


92.9 


102.3 


100.0 


96.1 



C-5 



-112- 



TABLE C-5 . Trends an Mean Enrollment for Advanced Degrees in Health Profes- 
sions by Carnegie Commission Categories of Institutions 3 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



49 


41 


43 


48 


50 


57 


80 


89 


19 


41 


46 


46 


52 


61 


101 


109 


12 


73 


70 


90 


92 


101 


110 


124 


18 


19 


21 


23 


21 


22 


37 


45 


71 


46 


50 


56 


65 


71 


103 


113 


21 


128 


139 


159 


182 


200 


251 


264 


21 


22 


24 


25 


29 


33 


60 


76 


28 


1 


2 


2 


3 


4 


25 


28 



Source: National Center for Education Statistics. 



TABLE C-5A . Trends in Index Numbers for the Means of Enrollment for Advanced 
Degrees in Health Professions Shown in TABLE C-5 (Means for FY 
1972 ='l00) 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 



1970 



1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



49 


51.2 


53.5 


60.7 


62.9 


70.8 


100.0 


112.0 


19 


40.5 


45.6 


45.5 


50.9 


60.3 


100.0 


107.8 


12 


66.3 


63.8 


81.8 


83.5 


91.6 


100.0 


112.9 


18 


51.9 


55.9 


62.6 


56.7 


60.0 


100.0 


122.7 


71 


44.2 


48.1 


54.3 


62.8 


69.2 


100.0 


110.0 


21 


51.1 


55.5 


63.5 


72.8 


79.7 


100.0 


105.4 


22 


36.8 


39.7 


41.8 


48.8 


54.7 


100.0 


127.8 


28 


5.2 


7.3 


6.6 


12.4 


16.3 


100.0 


112.7 



C-6 



-113- 



TABLE C-6 . Trends in Mean Enrollment for Advanced Degrees in Psychology by 
Carnegie Commission Categories of Institutions 



Type of 
Institution 



Number 



Fall Term of Fiscal Year 



1967 1968 1969 1970 1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



49 


63 


65 


66 


68 


70 


69 


76 


19 


80 


77 


78 


79 


81 


80 


85 


12 


72 


74 


76 


79 


80 


82 


92 


18 


39 


47 


46 


48 


53 


48 


55 


71 


79 


88 


100 


103 


100 


103 


108 


21 


131 


137 


158 


161 


157 


153 


164 


21 


79 


89 


84 


95 


97 


102 


104 


28 


40 


49 


69 


65 


61 


66 


68 



Source: National Center for Education Statistics. 



TABLE C-6A . Trends in Index Numbers for the Means of Enrollment for Advanced 

Degrees in Psychology Shown in TABLE C-6 (Means for FY 1972 = 100) 



Type of 
Institution 



Number 



1967 1968 



Fall Term of Fiscal Year 



1969 



1970 1971 1972 1973 



All Private Institutions 
Research Universities I 
Research Universities II 
Other Categories 

All Public Institutions 
Research Universities I 
Research Universities II 
Other Categories 



49 


91.5 


94.9 


95.4 


98.1 


102.2 


100.0 


109.9 


19 


99.7 


95.6 


96.7 


98.2 


100.7 


100.0 


105.5 


12 


87.5 


90.4 


92.9 


96.3 


97.6 


100.0 


112.6 


18 


81.6 


98.6 


96.0 


99.9 


110.2 


100.0 


114.5 


71 


76.7 


85.0 


96.9 


99.6 


97.3 


100.0 


104.5 


12 


85.6 


89.4 


102.9 


105.1 


102.1 


100.0 


106.8 


22 


77.4 


87.3 


82.5 


93.4 


94.7 


100.0 


102.5 


28 


60.3 


74.4 


104.0 


97.6 


92.0 


100.0 


103.1 



C-7 



-114- 



APPENDIX D 



TABLES SHOWING EARNED DOCTORAL DEGREES FOR INSTITUTIONS CLASSIFIED 
BY CATEGORIES OF THE CARNEGIE COMMISSION ON HIGHER EDUCATION 



TABLE D-l. Trends in Mean Number of Earned Doctoral Degrees in All Fields 
by Carnegie Commission Categories of Institutions 

TABLE D-1A. Trends in Index Numbers for the Means of Earned Doctoral Degrees 
in All Fields Shown in TABLE D-l (Means for FY 1972 = 100) 

TABLE D-2. Trends in Mean Number of Earned Doctoral Degrees in Biomedical- 
Behavioral Sciences by Carnegie Commission Categories of Insti- 
tutions 

TABLE D-2A. Trends in Index Numbers for the Means of Earned Doctoral Degrees 
in Biomedical-Behavioral Sciences Shown in TABLE D-2 (Means for 
FY 1972 = 100) 

TABLE D-3. Trends in Mean Number of Earned Doctoral Degrees in Biological 
Sciences by Carnegie Commission Categories of Institutions 

TABLE D-3A. Trends in Index Numbers for the Means of Earned Doctoral Degrees 
in Biological Sciences Shown in TABLE D-3 (Means for FY 1972 
= 100) 

TABLE D-4. Trends in Mean Number of Earned Doctoral Degrees in Health 

Professions by Carnegie Commission Categories of Institutions 

TABLE D-4A. Trends in Index Numbers for the Means of Earned Doctoral Degrees 
in Health Professions Shown in TABLE D-4 (Means for FY 1972 
= 100) 

TABLE D-5 . Trends in Mean Number of Earned Doctoral Degrees in Psychology 
by Carnegie Commission Categories of Institutions 

TABLE D-5A. Trends in Index Numbers for the Means of Earned Doctoral Degrees 
in Psychology Shown in TABLE D-5 (Means for FY 1972 = 100) 



D-l 



-115- 



in 
TJ 

rH 

4> 

•rH 

i— I 
rH 

< 

c 

■H 
W 

mo) 
m in 
r. G 
bo o 
oi -h 

*-> 
3 

rH <-> 

01 -H 
rl +J 
O W 

•p e 

o M 

o 

Q4H 
O 

Tj 

o W 
C 4) 
tn tH 
0) (H 

W O 
CO 

<4H CD 

O +J 

0) 

u u 

4> 

J3 B 
S O 
3-h 
Z m 
in 

e -h 

« g 

i» £ 
S O 
u 
e 

■H 0) 
■ H 

to 00 
XI 4> 
B B 
41 (h 

tH OJ 



OJ 

M 

< 



(H 

4> 
J3 
S 
3 
Z 



B 

o 

<4H 'H 
O +J 

3 

4) 4-> 
B.-H 

H in 

B 



OTtOM 

ohti eo 

CM tO rH 



O cm rt oo 

NIOH 



OHiOO 

ONio r^ 



Ol rH "» P> 



OlOlMh- 
00 Ol «3" >0 
rH CM rH 



(O CM o oi 

t^CONlO 
HNH 



t— Ol lO tO 
LD l/) O LT) 

HNH 



u-i cm cm oi 

"3- *3- Ol «3- 

rH CM 



t- f o> o 

NHN^t 
rH CM 



CM 00 O 00 



tO o rH rH 
r— tO «3* CM 



mOH« 
>»NHH 



in 

fiiam 
o CD <u 

•H -H -H 
+J +J +J 

3 -H 'H 
|J W in in 
• H rl rl 4> 
+J 4> 4> -rt 

W > > r. 

C -H -H O 

rH G B OO 

3 3 4) 

4) 4-> 

<->.B.e oj 

co o u u 
> rt rt 

•HKKI. 

i_ m o D 
Oli in in ,S 

)i«if 

rH PS OS O 



Ol OO tO 
rH r-Ol Ol 
CM tOrH 



0^00 H 
CM 00 Ol Ol 
NK1H 



rH rHO r- 
CM COOl 00 
CM K) rH 



moiOH 
rH r-~ 00 00 
CM tO rH 



CM tO LA CM 

o oo r- 

CM tO rH 



I- OliO rH 
I-. rH «3" \0 
i-H r-lrH 



to i-h o r~- 

1/1 00 CM "3- 

HISH 



00 CM CM \D 

lOOHK) 



Ol CO \0 Ol 
rH CM Ol CM 
rH CM 



O tj- rH in 

r-fuiH 



ONKlrt 
lON^H 



C— Ol O 00 

I/1HNH 



in in in 
B 4) 4) 

O -H -H 
•H +J 4J 
+J -H tH 

3 m in in 

u (■ mi 

■H 41 4> "H 

•M > > r. 

in -h >h o 
B w 



55 



OJ3.B 


CO 


•H 


u 


o u 


r-» 


rt 


tH 




-O 


CO 


CO 


r. 


3 


I) 


0) 


CD 


p. 


« 


w ja 




0) 


<u 


4H 


rH 


OS OS o 



CO 
•(-> 
to 

B 
o 

■H 
■M 

0! 
U 

3 

T3 

W 

rl 

o 
m 

tH 

0) 

•u 

B 
4> 

o 



CO 

z 



0> 

o 

rl 

3 

o 

CO 

01 



c 
o 
to 

in 

13 



in 
4) 

4) 
I- 
00 
CI 

Q 



ol 

rl 
O 
<-> 

u 
o 

o 

13 
0) 

B 

rl 

m 

w 

4H 
O 

m 

B 

35 o 
4>~ 
•C II 



Mr- 
Ooi 

•MrH 

£* 

™ CU 

5s 



BrH 
•H I 

a 
in 

"2 w 

B j 

<"5 



CO 
lO 
Ol 



OOOOtO CMCMrHsO 



O t- to CJ> 
O Ol O O 



Ol t- 0»0 

Ol Ol OO 



mom© to oo ^ rn 



to cm r— t— 

o o o o 



NOT* 

o o o o 



oooo oooo 



o o o o 
oooo 



oooo 
oooo 



"* m o «t 



oi f- r- oi 

Ol Ol O Ol 



r— Oi m to 

Ol oi Ol Ol 



tO ^- r1\D 



TKIlflCO rH m t— CM 
O1O1O00 Ol Ol coco 



vC 00 CM CM 



vo t— oo r- 

00 CO 00 I— 



o to \0 o 
oo oo r— c— 



M/IL/IO HN OT 



oo o r- oi 
t— oo r— o 



OO CM CM Ol 



Ol IO vOtO 

id r— \o m 



cm m oo * 
r— r— i© vo 



CM 00 Ol rH 

OOl/l* 



vOvOrHOO COOscCM 

tOiOOOCM toooto 
soomm toomto 



rl 

4) 

E 
3 



B 

o 

HH -H 
O +J 

3 

41 4-> 

P.-H 

H m 

B 



rH Ol O tO 



1-HCMvOO rH r— O r- 

TTtOtO tOtOCMrH 



rH rH 00 «cf 



r- o o oo 

tO * rO CM 



m O rH "3- 

ITNHH 



fi in in 

o oj o 

•H 'H 'H 
tJPH 
3 -H .H 

+j m in m 

•H rl (H 0> 

■M 4) 4) 'H 

m > > ri 

(Z'rt -H O 

rH C B 00 

3 3 4) 

4) V 

•l-> A A CO 

CO U O CJ 
> H rl 

CO CO ri 

4> V 4) 

in in j3 

4) 4) 4-> 

rH OS OS O 



t— CM CM ^t 

CM tO CM rH 



in in in 

C CD CD 
O "H 'H 



3 m in in 

+J rl rl 4) 
•H 41 CD -H 



u 

0- 



■M 


> 


> 


rl 


VI 


• rt 


■rt 





c 


B 


c 


ao 


hH 


3 3 


CD 








+J 


o.c 


CO 


•H 


<J 


o u 


rH 


rt 


rl 




JO 


CO 


CO 


rt 


3 


CD 


4) 


OJ 


tx 


in 


inx: 




4) 


4) 


■l-J 


f-H 


os oso 



D-2 



-116- 



> 

CO 
ft 
OJ 
03 



91 


01 


B 


C 








■H 


■H 


cc 


4-1 




3 


e 


w 


■H 


■H 




4-1 


0) 


0) 


01 


c 


id 


1-1 


u 




00 »H 


3) 


O 


a 






co 


rH 


01 


rr, 


•H 


H 


>-< 


O 





+-i 


01 


o 


OJ 


n 


4-1 


o 


cn 




c J 


■a 




CI) 


c 


C 


o 


r- 


■H 


CC 


K 





h 




U 


s-4 




(11 


01 


Xi 


•H 


fc 


DO 


3 


01 


Z 


c 




M 


CC 


01 


te 


U 


01 




S 


>, 




JJ 


c 




-rH 


m 




01 


in 


c; 


■n 


C 


c 


a) 


hi 


•H 


M 


') 


H 


CO 



h oo cs n r«jor^o 



O f* -$- 

CO ^* CN 



cy» m cn 
cn ** cn 



<r rsco cm 






0) 

| 

3 
Z 



oma \C o co -^ m o\ 

N -I « H -3- r-~ CO rH 



cm *3 cm -3- r^ co rH 



in cr> cm oo co co co m 

CM CO CM CO \D CO H 



*cr r- t— t r*- H vO m rj 
cm co cm CO m CM rH 



CM-d-r^co co O co o 

CM CO t-t CM UO CM 



CO O -4" IA lit OV CO vO 
rH CO rH CM -tf rH 



ps ic in in cn in in \o 

i-t CM rH CM *3 rH 



r- CO cm -3 o O -3 m 

iH CM rH CM -tf rH 



inOrH<- S Ci O CO 

^N HH inHM H 



c 
o 

4-1 -H 
O »-> 

3 

0) 4J 

O. -H 

>-, 4J 

H CO 

c 



re 


01 


01 




01 


01 


CO 







(11 


CD 




c 


0) 


0) 




•H 


•H 


•H 




CI 


•H 


TH 






4-1 


U 




•H 


4J 


4H 




S 


•H 


•H 




4-1 


■H 


•H 




4J 


en 


cn 


cn 


3 


01 


01 


0) 


•H 


u 


M 


ni 


4-1 


rl 


M 


■1) 




01 


1) 


■H 


iH 


CI) 


ai 


•H 


«: 


> 


? 


IH 


4-1 


> 


> 


!-i 


c 


•H 


•H 


r> 


CO 


—I 


•H 





►H 


c 


C 


mi 


c 


c 


(3 


an 




p 


3 


0) 


!-H 


C-J 


P 


01 


.1) 






4J 








u 


4-1 


X! 


£ 


CO 


o 


X! 


J3 


m 


m 


o 


I) 


U 


•H 


u 


CJ 


u 


> 


u 


l-i 




rH 


H 


M 




•H 


m 


(T 


u 


XI 


CO 


CC! 


M 


1-1 


n> 


01 


CU 


3 


01 


CD 


OJ 


p-4 


to 


01 


JC 


0- 


01 


0) 


.c 




m 


(II 


4H 




'1) 


(1) 


4J 


rH 


b: 


« 


o 


rH 


p; 


Pi 


o 



3 

m 



a! 



> 

X. 



01 




01 




0) 




1-1 




ou 




01 




Q 




rH 




CO 




1-4 




ci 


^— s 


4-1 


O 


(1 


c 1 


" 


rH 


a 






II 


•n 




ni 


CM 


c 


r*. 


M 


cr* 


CO 


rH 


W 






^ 


(4^ 


u- 









1-1 


01 


o 


n 


'4-4 


crt 




dl 


01 


£ 


c 




CO 


01 


(11 


-C 


S 


4-J 




1-4 


r-i 


CI 


1 


*H 


P 


CO 


W 


H 


►J 


0) 
.0 


9 


fci 


H 


3 




Z 


c 




■H 


X 




01 

•n 


s 


c 





rH 


-CC 



(0 

(1) 




>H 


CTn 




vO 




O* 


() 


rH 


01 




■H 




h, 










vC 




ON 



01 



co 



c 
o 

O 4-1 

CU 4-» 

O- -H 
>. 4J 

H cn 



cm <r in cn 



f^. \£> CN <X) 
O O O «H 



cn r^. <r o 



ON CO i— 1 CN 
CT\ Q\ iH O 



cn >»o O r*-. 



m <■ o o 
o o <-i o 



o o o o 



o o o o 
o o o o 



■-H O r-- \D 

OOHCO 



<J\ 1^- CO r-H 



cn o in .— I 
CT\ o »h r^ 



•^ H CO ^) 



s r>- o CO 

CO CO O v© 



<r in r^ <r 



CN CN vO O 
00 00 CTi sO 



O O CN v£) 



cn co co co 



cn \o n cm 
o \o >o <■ 



00 < vO ^ 



CC CO On m 

in m >x> <r 



(DM sD N O 



in O rH -3" 

<t (N H H 



O r*» cm r^* 
o o\ o o 



o o o o 



o o o o 
o o o o 



O CO vO o> 
O CT\ O 0> 



oo rH co r^. 



cn cn \o cn 

Q> CTi CTv CO 



rH <r so -<r 



oo <r o o\ 
co co o r» 



m o o cn 



<t m rH 

r^. r^. vo 



•^r co <ji h 



■sT >D CO vO 
vO s£> sD <T 



rH \o d in 



co m <t cn 
in so m cn 



CN O ON rH 



o ^o o 
vO <r cn 



O SO rH rH 



so cn cn <J 
>3- m <r cn 



1^- CTs O CO 



c 


CO 


01 




01 


01 


01 




o 


(11 


OJ 




c 


01 


01 




■H 


■H 


iH 




n 


-H 


-H 




4-1 


4-J 


4-1 




iH 


4-1 


4H 




a 


•r4 


■H 




4-1 


•H 


•H 




4-1 


CO 


1)1 


01 


rj 


0) 


OJ 


OJ 


•H 


M 


u 


CI) 


4-1 


1-4 


H 


OJ 


4J 


11 


1) 


•H 


•H 


OJ 


OJ 


•H 


01 


> 


> 


r4 


4J 


> 


> 


1-4 


C 


iH 


iH 


o 


01 


•H 


•H 





H 


C 


c 


Ml 


c 


c 


C 


DC 




p 


: j 


01 


rH 


=1 


P 


(11 


01 






4H 








4J 


4J 


,C 


.c 


CO 


O 


XI 


XI 


CO 


m 


CJ 


o 


(.) 


•H 


o 


u 


u 


> 


rJ 


>h 




rH 


u 


|H 




'rH 


ccl 


ccl 


U 


XI 


W 


CC 


M 


1-4 


(1) 


01 


(1) 


3 


OJ 


OJ 


0J 


PH 


(0 


01 


JS 


pu 


01 


0J 


X! 




cu 


0) 


4-J 




OJ 


III 


4J 


rH 


Pi 


OS 


o 


rH 


PC 


c2 


o 



D-3 



-117- 



ctj 

o 

•H CO 

00 O) 

o c 

rH o 

O i-l 

■H u 



01 


" 


o> 


M 


CU 




u 


uj 


01 


O 


CD 




n 


in 




o) 


IH 


■H 


~ 


U 


r< 


.1 


o 


00 


u 


01 


o 


u 


n 


m 


Q 


cj 


■tj 


c 


01 


n 


c 


■H 


Vw 


in 


w 


VI 


w 


H 


Mr 


p 


o 


n 




i ) 


u 




cu 


o> 


a 


m 


H 


00 


3 


01 


S- 


cr. 




ri 


a 


m 


m 


I I 


■11 




E 


>n 




fl 


a 




■H 


01 




'OJ 


en 


I) 


o 


c 


3 


01 


OJ 


■H 


u 


(J 


H 


tvi 



u 
a) 

a 



c 
o 

•4-1 *H 

O 4J 

3 

0) U 

>, *J 

H CO 
C 



r-. 00 rH vO 



00 On rH r- 



00 O M o 



^ *x> cm m 

H CS H 



m <r o in 



iA ii~t CM <" 



rsj o oo co 



o r-^ r^- en 



O vO r-» en 



O* \D T^ CM 



m o «h -a- 
<r rg iH rH 



\OmHH 
M »* CN H 



00 Ov CM CM 
CM <T CM rH 



CO O CO CM 

M iflN H 



v0 r»» CM ON 
CM -<f CM 



<-i on vD r- 

N fO H 



m rH r-( -<r 



CM U~) O C^i 



r-» o\ o co 

m rIN H 



c 


01 


111 




01 


in 


01 







01 


CJ 




c 


cu 


01 




H 


•H 


■H 




o 


■H 


•H 




u 


4-1 


U 




°H 


u 


U 




3 


■H 


■H 




u 


•H 


i-l 




■J 


in 


111 


in 


3 


01 


01 


01 


H 


— 


u 


ni 


U 


rl 


M 


01 


u 


cu 


01 


■H 


iH 


01 


ii 


■H 


CO 


> 


P 


u 


1-1 


> 


> 


M 


C 


H 


H 


o 


in 


H 


•H 





-t 


3 


c 


00 


c 


c 


C 


Ml 




o 


» 


HI 


r-l 


-J 


-J 


01 


01 






w 








u 


p 


c 


c 


3 


U 


.3 


J3 


id 


.Tl 


CJ 


u 


■H 


a 


O 


1 1 


> 


-J 


M 




^ 


M 


l-l 




■H 


« 


HI 


u 


J3 


m 


to 


t. 


In 


01 


OJ 


01 


3 


01 


tl 


01 


Xi 


in 


in 


c 


Ph 


01 


01 


c 




ii) 


01 


4J 




'11 


01 


o 


-I 


£K 


M 


O 


M 


2= 


X 


o 



C/l 



3 

-a 
ti 



u 
u 
1-1 
C 

a) 
o 



o 

.c 



to 
cj 

■H 

00 

o 
1-1 
o 

•H 
CO 

c 



01 
01 
01 

u 
oo 

01 

o 





1-1 
o 
u 
o 
o 

Q 

■a 

ai 
C 

Vj 

ctj 

w 



to 


o 


3 


< 1 


111 


^1 


0) 




S 


II 


01 


CN 


X 


r^ 



o >■ 



01 


u 


1-1 


o 


01 


I4H 


a 




1 


in 


-i 


c 


a 


m 




01 


X 


s 


cu 


^ 


Tl 




a 


ro 


H 


1 




« 


3 




■H 


w 




J 


n 


§ 


3 


H 


OJ 




l-l 


C 





rH 


n 


CM 


O 


ro 


rH 


d CM 


iH 


CA 


ON 


H 

rH 


r~- 


On 


ON 


O CO 


o> 


ON 


^O 


>0 


O 


00 


<t 


o in 


vO 
















^ 


iH 


rH 


vO 


r*- 


ON 


r* 


o ON 


iH 


00 


CO 


CT\ 


vO 


00 


CO 


O r» 

rH 


CO 


~d- 


rn 


CO 


<-. 


CO 


^j 


CM <J- 


\D 
















S\ 


m 


LO 


CN 


<r 


<r 


f* 


u-i cm 


r-\ 


00 


aj 


rH 


o-i 


r^ 


r^ 


o. *o 



u 

e 

53 



in -j- cm o 



mD <x> r^ co 

CTv CJ* 00 O 



v£> CO r^ o 



*0 f^» \0 \D 



rH oo m vD 
CT\ CO ON CT« 



a\ <r r^. en 



vo r*- m <r 
o^. a* o co 



o o o o 



CO xD o r-* 
<J\ &\ O O 



o o o o 



o o o o 
o o o o 

rH l-t l-l r-t 



rH CM \0 O 



cm o in vo 

O O CM CO 



rH en oo o 



o o o o 

o o o o 



en en r-** o 



O l-* \0 i-H 

o ON o o 



■<r m cm o 



in m m o 



r- o m en 

\0 r** r* -d- 



vO vD oo o 



^OM/1 O 

in in v£> «^- 



vd m i— i o 



m m o on 
m m r*- en 



r^ so in o 



cm en r- o 

in m ^£> en 



in o <— t <r 

<r CM r-H rH 



O <■ o> o> 



m m m in 

vD vD O -nT 



O O CO \D 



oo en >>* -<r 

m *n u*i m 



<r o o <r 
in ^d in en 



rH m t-H <r 



r*» o> o co 

in rH CM iH 



3 


m 


01 




01 


01 


01 




n 


01 


01 




C 


01 


OJ 




■H 


■H 


■H 




O 


■H 


•H 




U 


W 


4J 




•rt 


U 


4-> 




3 


■H 


■H 




4-1 


•H 


•H 




u 


in 


m 


01 


3 


01 


0) 


01 


H 


)H 


>r 


01 


i-i 


M 


In 


01 


IH 


01 


01 


•H 


•H 


01 


HI 


■H 


m 


> 


> 


IH 


4J 


> 


> 


|H 


C 


•H 


•H 


o 


01 


■H 


-H 





rH 


3 


3 


01) 


C 


3 


3 


00 




tJ 


TJ 


0) 


M 


o 


33 


01 


(1) 






^-i 








u 


U 


.C 


J3 


m 


U 


X 


-3 


id 


a 


CJ 


U 


I J 


■H 


t) 


CJ 


c ) 


> 


In 


IH 




rH 


\H 


l-l 




H 


OJ 


CO 


rl 


43 


B 


m 


n 


ll 


0) 


01 


0) 


3 


CD 


0) 


01 


CH 


01 


0) 


ifi 


Ph 


01 


CD 


J3 




.1; 


1) 


4-1 




ill 


01 


4H 


•-< 


od 


a 


o 


rH 


X 


« 


o 



D-4 



-118- 



o 
u 



re 

0) 



1.1 


c 


01 





91 


•H 


U 


4J 


01 


3 


01 


4-i 


P 


1-1 




u 


H 


X 


re 


c 


i- 


H 







4-1 


UH 


(1 





n 




a 


0: 




01 


■n 


•H 


01 


Ii 


E 


o 


re 


0) 


U 


u 




re 


Uh 


w 









C 


1* 





(1) 


•H 


-o 


0] 


a 


01 


3 


■H 


Z 


| 


a 




! ) 


01 




g 


0) 




•H 


c 


oil 




o 




a 


01 


M 


•V 


m 


c 


i 1 


0) 





H XI 






111 u 

H CO 

C 



csicni-Ho ei vo ro o 



r-H CM rH r-1 mvOfHO 



H N HO n\C (MO 



(N N tN H CO ^D CM rH 



tH iH CM O NOHO 



HHNO N ^ HO 



iH rH O O CM -d- iH O 



HHHO tM -J N O 



rH rH O O NK1HO 



tH rH O CS-J H O 



OOOO NPlrtO 



«10H-» 
ST N H H 



NOlO W 
l/"l rH CM rH 



c 


0) 


on 




01 


in 


CO 




o 


01 


01 




c 


01 


01 




•H 


1-1 


■H 







•H 


■H 




i-l 


4-1 


VJ 




•H 


U 


4-1 




3 


■H 


•H 




-u 


•H 


■H 




+J 


en 


C/l 


01 


3 


01 


01 


rn 


•H 


H 


l-i 


01 


4J 


n 


M 


0) 


4-1 


•11 


111 


-h 


■H 


(11 


01 


•H 


01 


> 


> 


V4 


4-1 


> 


> 


r1 


c 


■H 


■H 


1 1 


0) 


— I 


•H 





H 


C 


c 


on 


e 


3 


C 


Ml 




P 


» 


0) 


H 


CD 


a 


01 


n 






u 








4-1 


4-» 


J3 


j= 


re 


CJ 


JS 


a 


re 


m 


O 


u 


CJ 


■H 


CJ 


U 


u 


> 


l- 


l-l 




rH 


M 


s-. 




•H 


re 


111 


u 


,0 


It) 


m 


l-I 


U 


01 


01 


•11 


3 


0) 


0) 


II) 


Oh 


en 


CO 


.a 


U- 


0) 


01 


-c 




01 


(1) 


4J 




01 


0) 


u 


rH 


nd 


Pi 


o 


H 


pj 


pd 


O 



c 

01 



re 
■z. 



u 

3 
O 
c/l 
CD 







re 

01 



01 

m 
u 

00 
01 
Q 





a 

■o 

01 

c 



CO 


n 


c 


6 


m 


H 


(ii 




£ 


II 


0J 


CN 


J3 


r^ 



O bH 



01 


1-. 


H 


CJ 


1) 


ih 


J3 




h 


01 


re 


C 


S3 


re 




(ii 


X 


a 


01 


» — 


Tl 




C 


St 


H 


1 




a 


c 




■H 


LU 




hJ 


01 


3 


C 


H 


III 




r-1 


c 


H 


■H 



co -o o co 



<■ oo r-» co 



co o O r^ 

■— , O m mo 

O O r* vo 

i-t <-t cm 



OOOO 



OOOO 
OOOO 



m \D N N 



D 
^ 



o 

M-t -r-i 
O 4J 

H W 

C 



<i m -^ vo 

O O Ul iD 

H h Hen 



H <■ r* en 



on o> ft <*"> 
co vo <■ m 



u-i 00 CO O 



m CN v£> o 

^£> m o 



o> r*- O O 



o vo m o 
in vo cn 



co r** o r-. 



H vOOO 

vo vo m o 



co o f) o 



rH O iH O 

•4- in m 



CO ON co o 



H COvO O 

■sT en in 



U3COCOO 



co r- co O 

N IMH 



in o iH <r 

<T Cm iH fH 



co cm o co 



r^ r^ ^o r»- 

O ON ^ N 



<T o o o 



O cm cm O 

ONH<0 



OOOO 



OOOO 
OOOO 

iH tH r-4 r-l 



<■ m o tH 



ON CO CO H 

on o r-. tH 



OCOO-J 



in cm <r <r 
co o m >3- 



n rv o <■ 



co co o <r 
\o r^ m -j 



r^. on o i— • 



CO nO CO H 
in IN cN H 



m co o O 



r- On O i— t 



CO 00 CM i—l 

p^ oo m tH 



m co O cm 



H <f CO (N 
UO vO CM CM 



in co o o 



r-l ON <]■ O 

m m sf 



r-» on O co 



c 


l/l 


CO 




CO 


01 


CO 







0; 


01 




c 


HI 


01 




■H 


•H 


■H 




<) 


-H 


•H 




4-J 


U 


u 




■H 


1-1 


4-1 




re 


•H 


■H 




4J 


T-l 


•H 




4J 


in 


CO 


01 


re 


CO 


CO 


CO 


■H 


U 


u 


m 


4-1 


u 


l-I 


01 


1-J 


ai 


(1) 


•H 


•H 


01 


(U 


•H 


10 


> 


> 


H 


4J 


> 


> 


l-l 


C 


■H 


■H 





01 


•re 


•re 


n 


H 


C 


c 


an 


c 


c 


re 


HO 




P 


S3 


01 


M 


1=) 


B 


01 


01 






4-1 








vj 


u 


re 


a 


re 


O 


XI 


A 


re 


01 


o 


U 


o 


•H 


u 


o 


i_' 


> 


kj 


U 




r-l 


1J 


t- 




■H 


m 


Hi 


tj 


43 


re 


m 


l-i 


h 


01 


01 


oi 


re 


01 


01 


01 


ft. 


CO 


CO 


X. 


a. 


[0 


CO 


-C 




01 


01 


4-1 




m 


01 


4J 


iH 


c£ 


ai 


O 


r-l 


« 


oi 


o 



D-5 



-119- 



>N 

60 

o 

rH 
O 

xt 
cj 
>n 

CO 



01 




0) CO 


at 


en 


ij 


e 


tic 


n 


01 


— i 


a 


u 




? 


t-i 


u 


CD 


■H 


M 


4_J 


O 


to 


^J 


C 


cj 


H 







P 


cm 




c> 


•a 




01 


CO 


c 


0) 


rH 


•H 


ccl 


n 


fcl 


a 




bo 


UH 


<u 


o 


i-j 




m 


In 


U 


01 




x; 


c 


B 





3 


•H 


z 


CO 



E 







1 1 


c 




•H 


0) 




•M 


cn 


BO 


TO 


CI) 


C 


c 


<u 


u 


u 


CC1 


l-< 


u 



0) 

■i 

Z 



C 

o 

M-J -H 
O U 

3 

HI 4-1 
O. -H 
>n -u 

£H CO 

C 



cn r^ cn m 



N N M 1/1 



O -CT O -3- 



O -a- O en 



o en i-i en 



o m o en 



oo h o> n 



CTN CN ON CO 



CO CN NO CN 



r-- o r-~ cn 



n cn in rj 



m o "H -3- 

^N rlH 



-3- on en oo 



CS CO H CO 



CN On ON 00 



cn 00 O r- 



r-i oo CO r-~ 



o»io* 



co cn co m 



oo en r~ -3- 



SO CN <T CM 



*D Hst N 



NO CN <- CN 



r~ on O oo 

U"l rH CN rH 



cn 
















c 


cn 


to 




to 


01 


01 







01 


0) 




c 


CU 


01 




■H 


■H 


-h 







-H 


-H 




4-1 


^j 


4-1 




•H 


4-J 


4-J 




3 


■H 


'rH 




4J 


-H 


•H 




4-J 


01 


oi 


to 


3 


cn 


CO 


01 


-H 


1-1 


i-i 


a 


4-1 


m 


Ih 


0) 


4-1 


01 


ill 


•H 


•H 


01 


.1) 


'H 


to 


> 


> 


4-1 


4-1 


> 


> 


>H 


c 


-H 


•H 


o 


to 


•rH 


-H 


cl 


1-. 


C 


c 


00 


c 


c 


C 


00 




S 


;-j 


cu 


l-l 


B 


^> 


0) 


0) 






4J ' 








4-1 


4-1 


x: 


.c 


C!t 


O 


X. 


X! 


Cfl 


eo 


i/ 


u 


O 


•r-l 


u 


CJ 


CJ 


> 


VJ 


i-i 




rH 


M 


lH 




•H 


cn 


cn 


l-l 


XI 


CO 


CC! 


>H 


t-> 


cu 


CU 


tu 


3 


(1) 


"J 


tt) 


IH 


01 


If, 


.n 


P- 


01 


01 


J3 




cu 


cu 


4-1 




(1) 


CJ 


4J 


>H 


a 


tc 


o 


rH 


a; 


Di 


o 



X) 



c 
cu 



CO 



o 

CO 

;>. 

oo 

o 

r-H 

O 

o 

>N 
CO 



0) 

cu 
u 

00 

cu 

Q 



CO 

)H 

o 
u 

CJ 

o 
o 



IH 

CO 
W 



CO >i 
lH (U 



cu 




XI 


u 


E 


C! 


~ 


C1H 


EC. 






01 


X 


c 


cu 


CO 



t-1 


1 




o 


r/i 




■n 


w 


C 


r-l 


0) 


3 


H 


H 



cm in >* r*- 



oo m CM \£> 



O o\ m a\ 

IN H (N H 



o o o o 



o o o o 
o o o o 



o> co r-. <r 



v£> ON CO i— I 



r* O O O 



00 vO r-» O 



ocoo^n 
CO rs co I s - 



o> o rs vD 



■i 



O 4J 

3 

0) 4J 

& *H 

H CO 

C 



cm r-. a\ co 

c^ co oo CM 



o> f^> <r o 



r*- r* \o o 

n covo m 



f^ rs o a> 



^O vO H CO 

^o \D n in 



<r iH in i— i 



rH cm o r^ 

s co m m 



in o »h <r 

•^(NHH 



v£) O vO Oi 



m ^ h h 
«—. o -d- r-i 



CO \£> CT\ a> 



mo h co 

O O CM O 



o o o o 



o o o o 
o o o o 



cn o <r o 



H CO cm m 

O CT. H U\ 



H vO n CTi 



ON 


NO 


<r 


m 


U-N 


cn 


■3 


cn 


CC 


rH 


ON 


ON 


r-H 
rH 


r-. 


CTn 


ON 


ON 


'X) 


ON 


CO 


OO 


r^ 


O 


ON 


o 


r-o 


in 


NO 


















3N 


ON 


CN 


<r 


m 


CO 


cn 


nD 


rH 


rH 


ON 


O 


o 


r-s 


CO 


r^ 


r-H 


CO 



co <r \o o 



O <r r» cn 

N >£i CO \0 



ps (S I s - vO 



m r^ -»a" o> 
vO vo r*» *^- 



r-. oo on co 



<r in co co 

m ^ < n 



cm co -<r <j- 



o <r <■ cm 



OMNOm 

^- so <r cm 



("^ CT\ O CO 



to 


01 




CO 


01 


01 




a 


tu 




C 


01 


CD 




•H 


•H 




o 


■H 


•H 




4-1 


4J 




-H 


4H 


4J 




•H 


■H 




4J 


•H 


■H 




to 


01 


to 


3 


01 


01 


0) 


lH 


'H 


cu 


4H 


>H 


Ih 


CU 


0< 


01 


■H 


■H 


01 


OJ 


■H 


> 


> 


M 


4J 


> 


> 


lH 


•H 


■H 


n 


CO 


•H 


■H 





c 


C 


oo 


C 


C 


C 


on 


» 


B 


tu 

4-1 


M 


a 


3 


01 

4-1 


J3 


.CO 


en 


U 


,C 


X 


CO 


CJ 


u 


CJ 


•H 


u 


CJ 


U 


Ih 


lH 




rH 


rH 


M 




cn 


01 


|H 


J2 


m 


CO 


In 


CU 


0) 


OJ 


3 


01 


01 


0J 


cn 


01 


rC 


PM 


0) 


01 


Si 


cu 


CU 


4-1 




OJ 


OJ 


4J 


or; 


p-3 


o 


iH 


BS 


cs 


o 



D-6 



PART II 



Effects of Federal Funds Upon Selected Health-Related Disciplines 



David E. Drew and John G. Wirt 
with the assistance of Frederick W. Finnegan, Jr., 
Misako Fujisaki, and A. Lee Laniear 



The Rand Corporation 



PREFACE 



The work reported in this volume was performed as part of a Rand Corporation 
research effort carried out in conjunction with the American Council on Education 
and the Association of American Medical Colleges for the President's Biomedical 
Research Panel (under NIH contract number PP7533). The purpose of the total 
project was to trace the effects of federal health-related research funds on the 
nation's medical schools and universities. The work conducted under the contract 
has included both descriptive research and impact studies. The results of the re- 
search are reported in four documents: 

Medical Schools — Descriptive: T. E. Morgan and D. D. Jones, Trends and 
Dimensions of Biomedical and Behavioral Research Funding in Academic 
Medical Centers: 1964-1974, Association of American Medical Colleges, Jan- 
uary 1976. 

Medical Schools— Impact Study: A. P. Williams, G. M Carter, D. S. C. Chu, 
S. Coleman, A. P. Massell, C. R. Neu, R. Rasmussen, and W. Rogers, The 
Effect of Federal Biomedical Research Programs on Academic Medical Cen- 
ters, The Rand Corporation, R-1943-PBRP, January 1976. 

Research Universities — Descriptive: Lyle H. Lanier and Ivars Zageris, A 
Study of Financial and Educational Trends in Research Universities, with 
Special Reference to Federal Funding of Health-Related Research, American 
Council on Education, 1976. 

Research Universities — Impact Study: This report. 

Our specific goal was to assess the unique association between federal funding 
of selected university science departments and various indicators of the viability of 
those departments. While our analyses grew out of significant national policy con- 
cerns and are relevant to those concerns, we have avoided drawing normative im- 
plications or issuing policy recommendations. Our hope is that these findings will 
be useful to life scientists, researchers who study graduate education, and policy- 
makers concerned about the strength of the nation's life and behavioral science 
departments. 



m 



SUMMARY 



Rand's study of the effects of federal funds upon the viability of selected biomedi- 
cal and behavioral disciplines was based on multivariate analyses of a longitudinal 
data base, supplemented by case study field trips to nine carefully selected universi- 
ties (plus one pilot visit). 

In the quantitative analyses, we constructed models to predict several indices 
of academic department structure and function — e.g., Ph.D. production. These 
models included factors from earlier years that might have affected the variable 
under study. In each case, the key independent variable was federal expenditures 
for the department. Through this approach we were able to isolate the unique 
association of federal funding with each of these indices while controlling for other 
factors that also affected them. 

Perhaps the most important quantitative indicator of departmental structure is 
faculty size. In addition, several key qualitative aspects of departmental structure 
were relevant. The thrust of research in the biological sciences has changed dramati- 
cally over the past 15 years in the direction of molecular and cellular research. In 
our field trips we examined the shifting structure of life science departments as it 
related to federal funding. Similarly, through the field trips we explored the re- 
search productivity of the faculties. 

Many would argue that the reason for graduate education is the education of 
graduate students; these students also provide a valuable resource for faculty en- 
gaged in scholarly research. We examined the effects of federal funding upon gradu- 
ate enrollment and upon the production of the Ph.D. Much of the recent controversy 
about graduate education has grown out of the highly publicized unemployment and 
underemployment of Ph.D.s in the early 1970s. 

Although our analysis examined the separate predictors of Ph.D. production, 
graduate enrollment, and department size, we postulated that they were linked in 
a recursive model. 

We also wanted to assess the degree to which this federal funding effect varied as 
a function of the kind of institution. We compared funding effects in several catego- 
ries of universities. Any examination of higher education must differentiate public 
universities from private institutions. The history, budgeting, and orientation of 
these two sets of institutions are totally different. The relationship between federal 
funding and development of the basic life sciences at a university clearly differs 
depending on whether the institution has a medical school. In fact, the entire 
budgetary structure of the institution is greatly affected by the presence or absence 
of a medical school and the presence or absence of a hospital. We also examined the 
distinctive phenomena associated with the establishment of a new medical school 
at an institution during the past decade. 

A major theory about graduate education during the past few years held that 
leading departments reduced their enrollments and Ph.D. production as a result of 
federal funding trends while Cesser schools continued to produce large numbers of 
doctorates, the net effect being a reduction in the quality of the Ph.D.s in the country. 
This is a testable hypothesis. 



VI 



Our multivariate analyses are presented in numerous statistical tables. The key 
findings: 

1. Federal funding has a clear, strong, and positive relationship to depart- 
ment structure and function. This holds true even when other factors in 
our model are taken into account. 

2. A notable exception is graduate enrollment in psychology, which does not 
appear to be associated with federal funding. However, this exception prob- 
ably results from two factors: the large number of masters degree candi- 
dates in psychology and the combining of masters and PhD. candidates in 
the available federal data. 

3. Examination of the findings for public and private universities yielded a 
mixed profile. 

Graduate enrollments in biology consistently were more closely tied to 
federal funds in the public than in the private sector. 

Ph.D. production in both fields showed a stronger tie to federal funds in 
the private sector during the early years and in the public sector during 
the later years. 

In the more recent years, faculty size has been more associated with 
federal funds in private institutions than at the state schools. 

4. The presence or absence of a medical school had no consistent relationship 
to the effects of federal funds. However, our analysis of this factor was 
greatly hampered by limitations in available federal data. It was impos- 
sible to isolate, for example, graduate enrollments in a basic science at the 
main campus from those at the medical school. 

5. The effects of federal funds on faculty size and enrollment did not vary 
substantially between leading research universities and doctoral granting 
institutions; the results for Ph.D. production were mixed. 

6. Although a causal chain is difficult to establish, the level of federal funding 
is correlated with the level of other outside research support. 

7. Our analyses showed that other outside research funding had a smaller 
effect on these indices than federal funds. 

Carrying out this study has permitted us to make some observations about the 
problems and potential of research commissioned by the federal government to 
provide information on policy related issues. 

The government retrieves vast quantities of data each year from the nation's 
universities for research purposes. Yet, when a study like this one is to be conducted, 
information from separate agencies must be combined in a process that turns out 
to be quite difficult. It has been our observation that the data assembled by the 
different agencies vary considerably in their quality, that some are of poor quality 
and that little thought has been given to the process of merging these separate files 
for policy analysis. Thus, when meaningful questions are to be asked of the data, one 
discovers that two agencies have dealt with branch campuses differently, that disci- 
plines and professional fields are defined inconsistently, etc., substantially impeding 
the research process and greatly reducing the value of these data to policymakers. 
In our judgment, wiser interagency planning could reduce the number of requests 
for data to universities while greatly increasing the potential value of the data that 
are collected. 



ACKNOWLEDGMENTS 



Fred Finnegan and Misako Fujisaki played key roles in developing our comput- 
erized data base, maintaining it and carrying out the multivariate analyses we 
requested. Lee Laniear assisted us in conducting the field trips and in drawing up 
our observations about those visits. William Rogers carried out the computerized 
analyses that we used to select the field trip sample. 

Our colleagues from the Rand medical school impact study, particularly Albert 
P. Williams, Principal Investigator for the entire Rand effort, contributed numerous 
comments and constructive criticisms. 

Our colleagues at the American Council on Education, under the direction of 
Lyle Lanier, assembled most of the data we used in our analyses. 

A number of distinguished scientists and administrators served as advisors on 
our field trips; they are listed in Table 2. In addition we are indebted to the people 
at each university we visited who assisted our research effort. 

Finally, we received useful comments and suggestions from the Study Advisory 
Panel, chaired by Dr. Steven Muller, and from the formal reviewers of this report, 
Rand colleague Walter Baer and Donald Stewart of the University of Pennsylvania. 



vii 



CONTENTS 



PREFACE iii 

SUMMARY v 

ACKNOWLEDGMENTS vii 

Section 

I. INTRODUCTION 1 

Indices of Departmental Functioning 2 

Financial Factors Affecting Universities 2 

Components of the Departmental Budget 3 

Basic Structural Characteristics 3 

Central Issues Examined in This Research 4 

II. METHODS 5 

Data for Quantitative Analyses 5 

Field Trips 7 

III. RESULTS: QUANTITATIVE ANALYSIS 11 

A Model of Departmental Functioning 11 

Ph.D. Production 11 

Enrollments 15 

Department Size 17 

Comparison of Different Types of Institutions 19 

Cumulative Effects of Federal Funding 24 

Relationships Between Federal Support and Other 

Sources of Departmental Income 24 

IV. FIELD TRIP RESULTS 29 

University Administration 29 

Indirect Costs 31 

Financial Condition of Universities 32 

Organization of the Life Sciences in Universities 34 

Federal Funding and Faculty 36 

Federal Funding and Students 39 

Federal Funding and Outside Funding 41 

Science Policy Issues 41 

Uncertainty in Federal Policy 42 



IX 



I. INTRODUCTION 



The past two decades have been tumultuous ones for life science departments 
in this nation's research universities. The advances in our understanding of funda- 
mental biological phenomena have been spectacular, paralleling in significance the 
earlier revolution in physics. Behavioral and biological science departments in uni- 
versities have significantly increased in size during this period. Much of this growth 
has been supported by the federal government. 

The general pattern of support for research over the last 30 years has been well 
documented. Spurred by the spectacular success of the Manhattan Project during 
World War II, federal obligations for basic research increased dramatically through- 
out the 1950s and 1960s. These increases reached a peak in 1967 and then, for a 
variety of reasons, including an economy strained by the Vietnam War, obligations 
actually declined by a small amount over the next three years. In 1970, these federal 
commitments for basic research began to rise again, but at a much slower rate of 
increase than in the 1960s. 1 

Early in the deliberations of the President's Biomedical Research Panel, for 
which this study was conducted, a major concern was expressed about the viability 
of the nation's university life science departments in view of the ebbing of federal 
support. It is clear that the federal investment in scientific research at universities 
is not the result of a well-coordinated long range plan with widely understood 
parameters, but rather a mixed bag of programs developed for a variety of reasons 
and subject to abrupt increases, decreases, and cancellations. 

We wanted to see if the leveling of federal support had damaged university 
departments and their research activities. It is easy and perhaps misleading to cite 
examples of university activities that were briefly strengthened only to be weakened 
in the long run by federal support: a research lab standing empty (or not being fully 
utilized) because it was built under a huge grant that was not renewed; a public 
university totally out of balance because one area had been overextended when 
federal funds created faculty positions that were made permanent in the state 
budget to the detriment of other disciplines; and so forth. We proposed going beyond 
such anecdotal information. Our goal was to systematically trace the effects of 
federal funding upon a variety of indicators of departmental structure and function, 
whether those effects were positive or negative. This assessment required taking 
into account other sources of financial support of the departments as well as struc- 
tural characteristics of the university that were likely to affect the indicators we 
observed. Our basic goal was translated into a series of specific questions as outlined 
in the contract for this research. 

Below we shall discuss the indices of departmental functioning we used and the 
other components of the departmental budget (beyond federal funds) we considered. 
These elements were combined analytically in an organizational model that also 
took into account basic differences between types of universities. 

The methodology of the study consisted of two complementary approaches. The 

1 National Science Foundation, "Federal Funds for Research, Development and Other Scientific 
Activities," Surveys of Science Resources Series, Vol. 20 (NSF 71-35) and Vol. 23 (NSF 74-320), Washington, 
DC. 



main focus of the research was a set of multivariate analyses of longitudinal data 
about 148 of the nation's leading universities. However, in our judgment such math- 
ematical exercises conducted in a vacuum can sometimes lack a firm basis in reality. 
We therefore supplemented the multivariate statistical analyses with a series of 
field visits to a sample of nine carefully selected universities (plus one pilot visit). 
We do not pretend that ten short visits, no matter how well organized and intensive, 
can provide definitive answers to the questions we posed. However, they were ex- 
tremely useful in terms of generating hypotheses for the data analysis, testing 
relationships uncovered in those quantitative analyses, and providing some fairly 
well-grounded observations about the effects of federal funding upon the life 
sciences. 



INDICES OF DEPARTMENTAL FUNCTIONING 

Perhaps the single most important quantitative indicator of departmental struc- 
ture is faculty size. We were interested not only in trends in overall department size 
over time as it related to federal funding but also in the numbers of people at each 
professional level and, in particular, in the percentage of the faculty who were 
tenured. 

In addition to these quantitative indices, several key qualitative aspects of de- 
partmental structure were relevant to these issues. The thrust of research in the life 
sciences has changed dramatically over the past 15 years in the direction of molecu- 
lar and cellular research. In our field trips we examined the shifting structure of life 
science departments as it related to federal funding. Similarly, through the field 
trips we explored the research productivity of the faculties. 

Many would argue that the raison d' etre of graduate education is the education 
of graduate students. These students also provide a valuable resource for faculty 
engaged in scholarly research. In this study we examined the effects of federal 
funding upon graduate enrollment and upon the production of the Ph.D. Much of 
the recent controversy about graduate education has grown out of the highly publi- 
cized unemployment and underemployment of Ph.D.s in the early 1970s. There have 
been a number of assertions about the relationship of federal funding to enrollment 
and Ph.D. production at leading and lesser institutions. 



FINANCIAL FACTORS AFFECTING UNIVERSITIES 

This study was conducted against the background of the recent financial history 
of the nation's universities. Early in the 1970s, a number of observers described in 
rather stark terms the "new depression" in higher education. 2 The nation's higher 
education institutions, long accustomed to comfortable financial support from all 
sectors, were being forced to trim their sails for the first time in years. Although the 
bleakness of their financial picture may have been exaggerated at times, this clearly 
has been a much more difficult period financially for universities than the 1960s 
because of the coincidence of several factors: 

2 See, for example, Ear] F. Cheit, The New Depression in Higher Education, McGraw-Hill, New York, 
1971; and The New Depression in Higher Education— Two Years Later, McGraw-Hill, New York, 1973. 



Institutions that depend for a significant portion of their income upon 
donor gifts, notably the private universities, have suffered because of the 
decrease in philanthropic donations, which has been a by-product of the 
decline of the economy during the past few years. Even wealthy donors are 
less likely to give to universities because of the "relative deprivation effect" 
when they see the value of their portfolio reduced as the market falls. 
Parallel to the leveling of federal support has been a reduction in the 
enthusiasm and support state legislatures provide public universities. One 
reason was the reaction of a number of conservative legislators to student 
protests at some public institutions. 

State legislatures were also responding to the increased financial con- 
straints faced by states as the economy became weaker. With tight budgets, 
higher education had a difficult time competing with such items as feeding 
the poor and care for the aged. 

The highly publicized unemployment and underemployment of Ph.D.s 
made both national and state legislators reluctant to support graduate 
education. 



COMPONENTS OF THE DEPARTMENTAL BUDGET 

In developing our model of departmental functioning we took into account the 
many sources of income beyond federal support available to the sciences in a univer- 
sity. These included: 

• Other outside sources of support for research. For example, the American 
Cancer Society has supported the research efforts of young biologists in 
some states. 

• University support for "instruction and departmental research." 

• Federal support itself is not a unitary item. We were particularly interest- 
ed in the relationship of National Institutes of Health (NIH) funding to that 
from other sources. 

One of our major goals in this study was to investigate the relationships among 
these different sources of departmental income. Did an increase in federal support 
of a department tend to attract other sources of funds? 



BASIC STRUCTURAL CHARACTERISTICS 

Certain distinctions between types of universities were built into all our analy- 
ses. Any examination of higher education must differentiate public universities from 
private institutions because their history, budgeting, and orientation are different. 

The relationship between federal funding and development of the health-related 
sciences at a university clearly differs depending on whether the institution has a 
medical school. In fact, the entire budgetary structure of the institution is greatly 
affected by the presence or absence of a medical school and the presence or absence 
of a hospital. 

We felt it would be an oversimplification to treat these two dichotomies as 



representing absolute distinctions between totally different categories of schools. 
For example, some public institutions may be almost indistinguishable from private 
schools in many respects and vice versa. In short, this distinction is really not a 
dichotomy but a continuum. We assumed that both public and private institutions 
would show a full range of variation on the other characteristics we were examining 
— i.e., we made no assumptions about financial or departmental variables because 
we knew whether an institution was public or private. One practical effect was that 
we conducted all analyses for public institutions only, private institutions only, and 
for all schools combined. 



CENTRAL ISSUES EXAMINED IN THIS RESEARCH 

The general issues discussed above were presented as specific questions for 
purposes of the proposal to the President's Biomedical Research Panel and were 
then incorporated in the contract for this study. The specific questions we addressed 
in our work are listed below: 

A. Effects of Funding on Educational Programs 

1. What are the relationships between graduate enrollment in life science 
and behavioral science departments and federal biomedical funding? 

2. What are the relationships between federal biomedical (and related) 
funding and the rate of Ph.D. production by graduate life sciences and 
behavioral sciences departments? 

B. Effects of Faculty Hiring and Compensation Policy 

1. What are the relationships between characteristics of faculty in life 
science and behavioral science departments and federal biomedical 
and related funding? 

C. Changes in Funding 

1. What are the relationships between NIH/ADAMHA 3 funding to uni- 
versity departments and that from other sources? 

We attempted to go beyond merely answering these questions to some of the 
more general issues discussed above. For example, there has been little rigorous 
research with longitudinal data relating trends in federal funding to changes in 
faculty size, enrollment, and Ph.D. production in the life sciences. However, a major 
hypothesis about graduate education during the past few years held that as funding 
ebbed leading departments reduced their enrollments and Ph.D. production while 
lesser schools continued to produce large numbers of doctorates, the net effect being 
a reduction in the quality of the Ph.D.s in the country. This hypothesis is based on 
a number of testable assertions, none of which have been examined in the life 
sciences. Specifically, to what degree has Ph.D. production at either level been 
affected by federal funding? (Policymakers in Washington often act as though gradu- 
ate enrollments were strictly a function of federal support of students, and that if 
they cut back the support to students in a given field that will wipe out Ph.D. 
production in that field.) Our data provided an opportunity to test this assertion with 
multivariate techniques. 

3 Alcohol, Drug Abuse, and Mental Health Administration. 



II. METHODS 



As noted above, the methods for this study consisted of quantitative analyses of 
longitudinal data supplemented by field trips. 

During the field trips we were able to examine the effects of federal funding 
on the full range of departments representing the life sciences, especially biol- 
ogy, both at the main campuses and at related health professional schools. We 
also examined psychology departments and other social science departments. How- 
ever, in our quantitative analyses we were forced to draw some limits. Practical 
realities and theoretical considerations led us to reject the notion of analyses at the 
department level in favor of analyses at the discipline level. The biological sciences, 
in particular, are organized differently at every university. It would have been 
meaningless to analyze the relationship between federal funds to departments with 
identical titles — e.g., plant physiology — and, say, Ph.D. production in those depart- 
ments. Many of the people who study plant physiology might be housed in a depart- 
ment with a different name — e.g., Division of Biology. Thus, we decided to work at 
a higher level of aggregation, the discipline known as "biological sciences." A practi- 
cal factor in this decision was that many federal data items are collected only at the 
discipline level, presumably for the same reason. Some of the crucial variables we 
needed would not have been available at the department level. In the discussion of 
our quantitative work then, references to departments are, strictly speaking, refer- 
ences to disciplines. 

In the quantitative analyses, further decisions were made as a function of the 
structure of the available federal data. Our most important data source was the NSF 
survey on federal expenditures. In the survey the life sciences are divided into four 
categories: biological sciences, agriculture, clinical medicine, and miscellaneous. 
Psychology is a separate discipline, not included in the life sciences. We chose to 
conduct our analyses for the fields of psychology and the biological sciences. Data 
on agriculture were not available from some of the other federal sources, although, 
as noted below, the financial agriculture data were merged with those of the biologi- 
cal sciences in the data we received. Our focus upon university departments preclud- 
ed an interest in clinical medicine, which is located almost entirely in the medical 
school. In addition, the inclusion of this category within life sciences ruled out the 
possibility that we could conduct useful analyses of the higher level category "life 
sciences." 

In short, we analyzed data on "psychology" and "biological sciences" in our 
quantitative analyses; we visited the full range of life and behavioral science depart- 
ments in our field trips. 



DATA FOR QUANTITATIVE ANALYSES 

Longitudinal data on a series of indices of departmental structure and perfor- 
mance were deemed to be necessary for the quantitative analyses. Wherever possi- 
ble, we assembled data on the time period 1964-1975. To study the concepts discussed 



above, it was necessary to draw upon several federal data sources. The subjects and 
the federal data sources are indicated below: 

Departmental teaching staff: NSF 1 manpower survey 

Ph.D. production: NCES 2 survey of earned degrees conferred 

Graduate enrollment: NCES enrollment survey 

Federal expenditures: NSF expenditure survey 

Federal institutional obligations: NSF CASE' J survey 

NIH funding: NIH IMPAC file. 

Assembling, merging, and cleaning these data, largely the responsibility of the 
American Council on Education, 4 was a Herculean task. The effort taught us a great 
deal about the inadequacies of data available on national graduate education and 
the technical obstacles that must be overcome to be able to conduct meaningful 
analysis of graduate education. Among those technical problems: 

• There are inconsistencies from data source to data source in the definition 
of fields and disciplines; 

• Some of the surveys are not conducted on an annual basis; 

• Frequently the data gathered in early years was of poor quality; 

• Federal agencies differ considerably in their treatment of information from 
multi-campus systems; there are inconsistencies in terms of whether data 
are aggregated for the entire system or reported by campus, when new 
branch campuses are identified, and so forth. 

Several of the technical problems we encountered with these data had serious 
implications for the interpretation of our results. The National Science Foundation 
survey, which compiled information on federal expenditures, did not disaggregate 
expenditures in the biological sciences from those in agriculture during the earlier 
years. Consequently the figures for those two categories were combined in the finan- 
cial data made available to us. 5 However, information from other federal agencies — 
e.g., on graduate enrollments in biology from the Office of Education — did not in- 
clude data on agriculture. The net effect of this was to add noise to our analyses of 
the biological sciences, probably deflating the correlations observed there. This 
factor somewhat clouds our comparisons of the public and private sectors since 
federal expenditures for agriculture are for public institutions. 

The federal data files — e.g., on enrollments, Ph.D.s. and expenditures — fail to 
separate data from the medical school from those items obtained from the main 
campus. For example, all biochemistry Ph.D.s, whether at the medical school or the 
main campus, will be reported as one datum for a given university. This factor 
seriously confounds'comparisons of life science departments at universities without 

1 National Science Foundation. 

2 National Center for Educational Statistics of the Office of Education. 

3 Committee on Academic Science and Engineering. 

4 These data sources are described in detail in: Lyle H. Lanier and Ivars Zageris, A Study of Financial 
and Educational Trends in Research Universities, with Special Reference to Federal Funding of Health- 
Related Research, American Council on Education, 1976. 

5 The departments included by NSF in each category are: Biological Sciences (anatomy, biochemistry, 
biophysics, biogeography, ecology, embryology, entomology, genetics, immunology, microbiology, nutri- 
tion, parasitology, pathology, pharmacology, physical anthropology, physiology, botany, zoology,) and 
Agriculture (agricultural chemistry, agronomy, animal science, conservation, dairy science, plant 
science, range science, wildlife). 



medical schools and main campus departments of universities that happen also to 
have medical schools. 



FIELD TRIPS 

The ten universities that we visited were selected to be a representative sample 
of the entire range of institutions receiving federal health-related research funds. 
Our primary criterion in selecting the sample was the total amount of federal R&D 
funds received by the university in FY 1971, which was the midpoint of the decade 
we were studying. This criterion was chosen so that we could compare the effects 
of recent changes in federal research policy on universities that have varying de- 
pendence on federal funds. The data were from the NSF CASE obligations survey. 

Four other criteria were chosen to distinguish among universities according to 
basic institutional characteristics. A key criterion was whether the university was 
a public or private institution, since we wanted to explore how these kinds of schools 
may be affected differently by changes in federal policy. 

A second criterion was whether the university had a medical school. Although 
this study was primarily concerned with life science departments in schools of arts 
and sciences, we wanted to explore possible ways that these departments are affected 
by the existence of a medical school in the same university. 

A third criterion was the quality of the university's faculty in the life sciences, 
as ranked by Roose and Andersen. 6 We wanted our sample of universities to include 
institutions ranging in quality from "medium" to the highest level, 7 so that we could 
explore the extent to which differences in quality are tied to federal funding. 

Finally, we wanted the sample of universities to be geographically distributed 
across the country. We defined four separate regions of the country (West, South, 
Central and East) and selected universities accordingly. 

The actual sample often universities was drawn using a finite sample selection 
algorithm developed by Carl Morris of The Rand Corporation. This algorithm ac- 
cepts both continuous and discrete variables as input to the selection of an efficient 
sample, where efficiency is measured in terms of the ability of the sample to separate 
the effect of the input variables on the characteristics to be studied. For example, 
one seeks to avoid a sample of five private universities without medical schools and 
five public universities with medical schools. The program accepts information on 
the relative importance of the input variables and allows constraints on the composi- 
tion of the sample. The sample drawn is superior to both random and representative 
samples because a random sample will occasionally confound the effects to be stud- 
ied, and both a random sample and a representative sample concentrate on the 
less interesting subpopulation of mediocre, poorly funded institutions. 

In selecting the sample for the field trips we specified the following criteria and 
conditions: 

G K. D. Roose and C. J. Andersen, A Rating of Graduate Programs, American Council on Education, 
Washington, D. C, 1970. 

7 We denned a university to be top-ranked if three or more of its biology departments were placed by 
Roose and Andersen in their distinguished category. We defined a university to be "medium" if two or 
less of its departments were in Roose and Andersen's top category but its departments fell in their middle 
category. Departments ranked by Roose and Andersen in the lowest sector (or unranked departments) 
tend not to receive federal research funds. 



Exactly five private and five public universities were to be selected, and 

university control was given very high weight. 

Whether the university had a medical school was a variable whose effect 

it was important to understand, so it was given a high weight. 

The level of federal funding was given a high weight, and and a parabolic 

term in federal funding was included so that the influence of funding would 

be observed at intermediate funding levels as well as extreme ones. 

Exactly six top-ranked and four middle-ranked universities were to be 

selected, and a medium weight was specified for faculty quality. 

At least one university was to be selected from each region of the country, 

with an informally imposed constraint that no region would be overrepre- 

sented. 

At least one university that had high faculty quality and a below-median 

level of federal funding was to be selected. 

At least one university that had been visited previously in the Rand study 

of medical schools was to be included. 

The sample had to include three schools that had previously been selected 

for early visits in this study. 

Table 1 indicates the characteristics of the institutions selected for our field trip 
sample. In the interests of confidentiality the names of the ten universities have 
been omitted. 

The field trips to universities were one to two day visits. Typically, the field trip 
team consisted of several members of the Rand project staff, an experienced higher 
education administrator, and a distinguished life scientist. 8 During the day, we 
conferred with key administrators at the university — e.g., the Vice President for 
Research, the President, the Graduate Dean — and interviewed department chair- 
men, senior faculty, junior faculty, and graduate students in those departments that 
received significant federal funds for health-related research. Interviews with ad- 
ministrators and scientists from medical schools and related health professional 
schools on the campus also were scheduled. 

Each site visitor was given a list comprising the key issues that were to be 
examined during the visit: 

1 . In what areas of science did the most growth occur in this university during 
the 1960s? Why? 

2. What had been the university's policies regarding the growth of the biologi- 
cal sciences with respect to priority areas of growth, faculty size, research 
productivity, types of new faculty sought? 

3. What growth actually occurred in the biological sciences? 

4. What was the role of federal funds in financing whatever development 
occurred in the biological sciences? 

5. What are the administrative mechanisms through which research dollars 
are incorporated in the departmental budget? 



8 The field trip consultants are listed in Table 2. Their comments aided us greatly in drawing together 
the general observations presented in Section IV. However, the authors assume full responsibility for the 
statements in that section. 

On our pilot field trip we joined forces with two colleagues, Bruce Smith of the Brookings Institution 
and Charles Kidd of the Association of American Universities, who are conducting a study of science in 
America's research universities. 



Table 1 
FIELD TRIP SAMPLE 



School 


Control 


Medical 
School 


Distinguished 
in Biology 


Total Annual 

R&D Funds 

FY71* 

(in thousands) 


1 


Public 


Yes 


No 




$10-15,000 


2 


Private 


Yes 


No 




10-15,000 


3 


Public 


No 


No 




< 5,000 


4 


Public 


Yes 


Yes 




> 35,000 


5 


Public 


No 


Yes 




10-15,000 


6 


Public 


Yes 


No 




10-15,000 


7 


Private 


No 


Yes 




10-15,000 


8 


Private 


Yes 


Yes 




20-25,000 


9 


Private 


No 


Yes 




< 5,000 


10 


Private 


No 


Yes 




15-20,000 



*These data indicate total federal obligations for R&D to the 
institution, including professional schools. 



Table 2 



FIELD TRIP CONSULTANTS 

H. Vasken Aposhian 

Department of Cell and Developmental Biology 

University of Arizona 

Glen F. Clanton 

Deputy Provost and Dean for Planning 

Vanderbilt University 

Hans Laufer 

The Biological Science Group 

University of Connecticut 

Louis Levin 

Retired 

Formerly National Science Foundation administrator 

David McBride 

Director, Office of Research & Project Administration 

University of Rochester 

John Millett 

Vice President 

Academy for Educational Development , Inc . 

Meredith Runner 

Department of Molecular, Cellular, and Developmental 

Biology 
University of Colorado 

Michael Useem 
Department of Sociology 
Boston University 



10 



6. What are, and have been, the trends in student enrollment in the biological 
sciences? How are these trends related to federal funding? 

a. Number of M.S. and Ph.D. students enrolled. 

b. Proportion supported by fellowships, research assistantships, teaching 
assistantships. 

c. Number of postdoctoral students. 

d. Undergraduate student enrollments — any consequences for the gradu- 
ate program. 

7. What proportion of the faculty is and has been tenured? 

8. What are the current sources of outside support for research in the biologi- 
cal sciences besides the federal government — state government, private 
sources, etc.? 

9. Are there any cost commitments to biology made by the university in 
earlier years that currently are creating budget problems? 

10. Are there any serious imbalances in the biological science programs that 
have been caused by shifts in the pattern of federal funds? What mecha- 
nisms does the university have for restoring the balance in those groups — 
e.g., proportion of tenured faculty, ratio of faculty to graduate students? 

11. Has the average time to the completion of the Ph.D. changed during the 
past ten years? 

12. Have there been changes in the types of federal funding — grants versus 
contracts, ratio of research support to traineeships and fellowships? 

13. What was the university policy regarding hard versus soft funding during 
the growth period? 

14. What has been the effect upon this institution of shifts in federal policy 
with respect to investigator-initiated versus targeted research? 

15. What observations can the faculty and administrators in this institution 
contribute about the current peer review controversy? 

16. What is the institutional attitude toward the present federal percentage of 
levels of, restrictions on, and administrative procedures relative to indirect 
cost reimbursements? 

17. How are indirect costs handled adminstratively within the institution? 

We found that we were able to answer some of these questions more fully than 
others. 



III. RESULTS: QUANTITATIVE ANALYSIS 



We began our analyses by examining the basic questions from the contract one 
by one. Our study of the relationship of federal funds to each dependent variable — 
e.g. Ph.D. production — required elaborating a multivariate model of the determi- 
nants of that phenomenon alone. 

In the work reported below we were conducting cross-sectional analyses aimed 
at uncovering the unique association between federal funding and each of the de- 
partmental indices. We did this for different sectors of higher education, such as 
public and private, and for different years during the past decade. However, these 
were not causal analyses, and causal inferences should be drawn from these results 
only with the utmost caution. For example, we report a significant association 
between federal funding of biology departments in 1974 and the rate of Ph.D. 
production by those departments. The regression coefficient indicates that, even 
when other factors affecting Ph.D. production are considered, $1 million of federal 
funding at a university was associated with approximately one extra Ph.D. produced 
during 1974. This means that those schools with extra money produced extra Ph.D.s, 
when the effects of other factors such as enrollment were controlled for. It does not 
necessarily mean that a school would produce one extra Ph.D. if it were to be given 
an additional $1 million in the forthcoming year. 

In short, our goal was to isolate the unique associations of federal funding with 
various departmental parameters in a cross-sectional analysis and to compare the 
strength of those associations in different types of institutions. 



A MODEL OF DEPARTMENTAL FUNCTIONING 

In addition to exploring predictors of Ph.D. production, graduate enrollment, 
and department size, we postulated that they were linked. An academic department, 
like any other complex system, consists of a network of interdependent structural 
units and functions. Through our analyses, which focused on federal funding, we 
hoped to elaborate and refine a model of these complex interactions. Figure 1 por- 
trays the network of relationships implicit in those multivariate analyses. Note that 
in addition to the direct effect federal funding has upon Ph.D. production, it has an 
indirect effect through its influence upon graduate enrollment and faculty size. 



Ph.D. PRODUCTION 

Perhaps no other indicator has had as much effect during this period of crisis 
for graduate education as the rate at which Ph.D.s are produced by the nation's 
graduate schools. In the mid-1960s, national support for graduate education was at 
its height, and most experts were predicting a shortage of Ph.D.s through the next 
decade. Economist Allan Cartter, however, took a different view. His analyses and 
projections indicated that the future demand for Ph.D.s would not be likely to exceed 
the supply. As a partial explanation of the differences between his conclusions and 

11 



12 



Federal $ 



Othe 



Ph.D. Production 




Teaching Staff 



"ig- 



■Proposed model of departmental functioning 



those of other specialists, Cartter noted that "educational researchers in govern- 
ment agencies had collected the wrong information for many years and had drawn 
hasty conclusions from imperfect data." 1 He was one of the first, if not the first, to 
predict an oversupply of Ph.D.s beginning in the 1970s. For example, in 1972 he 
stated: "We are on a course which would result in one-third too many Ph.D.s pro- 
duced in the latter part of this decade and perhaps one-half too many in the 1980s 
for the types of employment we have known in the past." 2 

In light of the accuracy of his early projections, Cartter's work has received a 
good deal of attention. But he is not without critics. For example, Vaughn and 
Sjoberg have considered some of Cartter's assumptions and have given a number of 
reasons for viewing his projections with some skepticism. They charge that Cartter 
"ignores fundamental social changes already underway within American society, 
changes that are likely to erode the very basis of his projections." 3 They argue that 
basic shifts in the nature of the American economy will lead a larger number of 
people to seek higher education than Cartter had assumed. The emerging primacy 
of the service sector in the economy implies a greater reliance on advanced educa- 
tion, as does the increase in leisure time and the growing demands of women, 
minorities, and others for advanced education on a part-time or full-time basis. 

While national policymakers have been responding slowly to this problem, a 
number of states have taken direct action. Foremost among these is New York, 
where the State Board of Regents is invested with considerable authority over both 
public and private education. Three years ago, a report recommended that the 
number of doctoral-producing programs be reduced, citing, among other factors, the 
overproduction of Ph.D.s by the state's higher education institutions. 4 A special 
study was commissioned to review and evaluate the adequacy of doctoral programs 
on a field-by-field basis, and to make recommendations about which ones should be 
abolished and which strengthened. 5 



1 Allan M. Cartter, "A New Look at the Supply of College Teachers," Educational Record. Vol. 46, 
Summer 1965, pp. 267-277. 

2 Allan M. Cartter, "Scientific Manpower of 1970-85," Science, April 9, 1972, p. 243. 

3 Ted R. Vaughan and Gideon Sjoberg, "The Politics of Projection: A Critique of Cartter's Analysis," 
Science, July 14, 1972, p. 142. 

4 "Meeting the Needs of Doctoral Education in New York State," New York Board of Regents, 
Commission on Doctoral Education, Albany, New York. January 1973. 

5 Decisions by the state to eliminate two doctoral programs at the State University of New York, 
Albany campus as a result of that study recently have been contested by SUNY trustees in a New York 
Supreme Court action. See "SUNY Challenges Right of State to Curb Courses," The Chronicle of Higher 
Education, March 8, 1976, p. 3. 



13 



The growth of graduate education during the 1960s occurred at different rates 
in different sectors of the academic world. For example, as Kidd has noted, over the 
decade doctoral production in the top 30 private universities dropped from 39 per- 
cent to 27 percent of the total of all doctorates produced, whereas in the public 
universities below the top 30, it increased from 9 percent to 24 percent. Overall Ph.D. 
production tripled between 1960 (10,000) and 1969 (30,000). Among the reasons Kidd 
notes for the differential growth rates of different types of institutions are the steep 
increases in state budgets for support of state institutions in the 1960s, the greater 
expansion of public universities in every aspect, and the pressure to provide teach- 
ing assistants for the rapidly growing undergraduate population at public universi- 
ties. 6 

Such observations as these about the differential growth rates of public and 
private institutions have led some observers to argue that Ph.D. output should be 
limited to elite institutions. 7 Ph.D. production is often seen to be an indication of 
graduate quality. For example, Cartter used a measure of doctorate production as 
the key criterion for inclusion of schools and programs to be assessed in his ratings; 
Roose and Andersen, in their replication of his study, followed this lead. 

To trace the effects of federal funds (F) upon Ph.D. production, we built a model 
specifying other departmental factors that might have a bearing on this index: 

Departmental graduate enrollment (E); 

Faculty size, in this case our measure of teaching staff (T); 

Control — i.e., whether the institution was public or private (C); 

Whether the institution contained a medical school (M); 

Other outside sources of support of the department, including private and 

state sources (O). 

The equation relating these variables to Ph.D. production (P) is shown below. 

P t =: B,E t .j + B 2 T t _ k + B 3 C + B 4 M + B 5 O t _, + B 6 F t . m . 

Prediction of Ph.D. production in year t is a function of each variable measured 
in that or a previous year. The lags for each variable were determined as follows. 
The zero-order Pearson product-moment correlation between a dependent variable 
at year t and the independent variable at year t, t — 1, t — 2, and t — 3 were examined 
and the year associated with the highest correlation selected. (This determination 
was based on the total sample; separate lags were not created for each subgroup — 
e.g., private institutions only.) 

In our analyses we constructed the above equation for each year from 1969 to 
1974. We examined the results to answer four questions: 

1. How successful was the model in accounting for Ph.D. production? In other 
words, what percent of the variance in this index was explained by the 
model? 

2. In a given year, what was the relative contribution of each of the indepen- 
dent variables in the model? 

6 Charles V. Kidd, "Shifts in Doctoral Output: History and Outlook," Science, February 9, 1973, pp. 
538-543. 

7 John R. Niland, "Allocation of Ph.D. Manpower in the Academic Labor Market," Industrial Rela- 
tions. Vol. II, No. 2, May 1972, pp. 141-156. 



14 



3. How significant was the association of federal funding, our key variable, 
with Ph.D. production? 

4. Did the importance of federal support vary across the different sectors of 
higher education? 

Table 3 indicates the success of the model in predicting Ph.D. production in the 
biological sciences; Table 4 presents similar information for psychology. The results 
are presented for all institutions, public only, private only, universities with medical 
schools, those without medical schools, and for each of the leading three categories 
of universities as defined by the Carnegie Commission on Higher Education. 8 



Table 3 



PREDICTION OF PH.D. PRODUCTION IN THE BIOLOGICAL SCIENCES 



(R 



Percentage of Variance Explained) 





N 1969 


1970 


1971 


1972 1973 


1974 


All Institutions 


148 


88 


.82 


.85 


.88 


90 


.85 


Public 


93 


95 


.85 


.91 


.91 


90 


.89 


Private 


55 


80 


.78 


.85 


.89 


90 


.74 


With Medical School 


68 


84 


.74 


.84 


.86 


90 


.79 


Without Medical School 


80 


92 


.89 


.88 


.92 


89 


.97 


Research Universities 1 


49 


87 


.77 


.81 


.85 


88 


.77 


Research Universities 2 


40 


88 


.67 


.72 


.87 


77 


.75 


Doctoral Granting 
















Universities 


47 


78 


.59 


.60 


.63 


61 


.72 



Table 4 



PREDICTION OF PH.D. PRODUCTION IN PSYCHOLOGY 
2 
(R = Percentage of Variance Explained) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 


148 


.84 


.80 


.66 


.62 


.66 


.65 


Public 


93 


.94 


.86 


.69 


.73 


.80 


.72 


Private 


55 


.78 


.61 


.61 


".39 


.50 


.58 


With Medical School 


68 


.83 


.76 


.64 


.58 


.58 


.59 


Without Medical School 


80 


.91 


.83 


.73 


.74 


.75 


.71 


Research Universities 1 


49 


.94 


.81 


.81 


.65 


.76 


.61 


Research Universities 2 


40 


.86 


.77 


.56 


.57 


.65 


.75 


Doctoral Granting 
















Universities 


47 


.83 


.71 


.49 


.45 


.64 


.57 



8 Statistical considerations — some missing observations combined with small sample sizes — led us to 
reject groupings based on combinations of the structural variables (e.g., private institutions with medical 
schools, private institutions without medical schools). 



15 



In biology the model accounted for 82 percent to 90 percent of the variance, 
depending on the year. The model was more successful in predicting Ph.D. produc- 
tion for public institutions than private, and for those without medical schools than 
those with medical schools. The R 2 tended to be somewhat higher for Research 
Universities 1 than for the lower two Carnegie categories. 

In psychology the model accounted for between 62 percent and 84 percent of the 
variance, depending on the year. Again it was more successful in universities that 
were public, had no medical school, and fell in the leading Carnegie category. 

We next examined the relative contribution of each of the variables in the 
model. The 1974 equation for each discipline based on all institutions is given below. 
The coefficients are standardized; those that are significant at the .05 level are 
indicated with an a and at the .01 level with a b. 

Biological Sciences: 

P = .81 b E - .05T + .08C" + .00M + .090 + .17 h F. 
Psychology: 

P = .64 b E - .06T + .04C + .11M - .040 + .32 h F. 

In both fields enrollment was the largest determinant of the rate of Ph.D. 
production. However, in both fields, federal funding was the next most important 
factor; in each case the coefficient for federal funding is significant at the .01 level. 9 

A comparison of the importance of federal funding across different types of 
universities with respect to Ph.D. production, enrollment, and faculty size is re- 
ported below. 



ENROLLMENTS 

A model of factors that were hypothesized to affect enrollment for advanced 
degrees — i.e., graduate enrollment (E) — was developed and used in multivariate 
analyses aimed at uncovering the unique effect of federal funds upon graduate 
enrollment. The following variables were included: 

• Faculty size — i.e., full-time teaching staff (T); 

• Whether the institution was public or private (C); . 

• Whether the institution has a medical school (M); 

• Federal expenditures in the discipline (F); 

• Other outside sources of support of the discipline (O). 

These variables were combined in a multiple regression model as follows: 

E t = BxTt.j + B 2 C + B 3 M + B 4 t k + B 5 F t _, 

As before, the lag for each independent variable was determined through inspec- 
tion of the Pearson product-moment correlations between the dependent and the 

9 The reader should note that the coefficient for federal funding indicates the direct effects of funding 
upon Ph.D. production, given a certain enrollment, faculty size, etc. The indirect effects of federal funding 
through the effect of funding upon enrollment or upon faculty size are not reflected in the coefficient for 
federal funding. Thus, the total funding effect upon Ph.D. production, both indirect and direct, is substan- 
tially larger than indicated by this coefficient. 



16 



independent variable measured in the same year and each of the three previous 
years; the year associated with the largest of the four zero-order correlations was 
selected. In order to explore the nature of the different relationships for public 
institutions versus private institutions, and those with medical schools versus those 
without, we ran separate regressions for each of those four subgroups. 

Tables 5 and 6 indicate the success of this model in explaining the variation in 
graduate enrollment in the biological sciences and psychology. In both fields, the 
explanatory power of the model is substantial but not as great as it had been for 
Ph.D. production. In both fields, enrollment can be predicted better at public institu- 
tions than at private. In biology the model does better at universities without a 
medical school and at those in the second Carnegie category. In psychology the 
model predicts enrollment better for universities with a medical school and those 
in the leading Carnegie category. For psychology the explanatory power of the model 
improves in the more recent years. 



Table 5 



PREDICTION OF GRADUATE ENROLLMENT IN THE BIOLOGICAL SCIENCES 



(R 



Percentage of Variance Explained) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 


148 


.57 


.59 


.57 


.53 


.56 


- 


Public 


93 


.63 


.67 


.64 


.66 


.74 


- 


Private 


55 


.57 


.57 


.62 


.54 


.48 


- 


With Medical School 


68 


.56 


.55 


.49 


.41 


.47 


_ 


Without Medical School 


HO 


.59 


.62 


.67 


.78 


.77 


- 


Research Universities 1 


49 


.35 


.42 


.45 


.43 


.42 


- 


Research Universities 2 


40 


.64 


.67 


.72 


.67 


.72 


- ' 


Doctoral Granting 
















Universities 


47 


.62 


.30 


.33 


.36 


.26 


— 



Table 6 



PREDICTION OF GRADUATE ENROLLMENT IN PSYCHOLOGY 
2 
(R = Percentage of Variance Explained) 



1969 1970 1971 



1972 



1973 1974 



All Institutions 


148 


.39 


.39 


.58 


.65 


.63 


- 


Public 


93 


.37 


.37 


.62 


.73 


.72 


_ 


Private 


55 


.36 


.50 


.39 


.35 


.31 


- 


With Medical School 


68 


.37 


.42 


.73 


.74 


.60 


- 


Without Medical School 


80 


.41 


.30 


.49 


.52 


.65 


- 


Research Universities 1 


49 


.51 


.53 


.86 


.79 


.76 


_ 


Research Universities 2 


40 


.16 


.13 


.60 


.49 


.41 


- 


Doctoral Granting 
















Universities 


47 


.54 


.41 


.34 


.54 


.45 


~ 



17 



The equations for the most recent year, 1973, generated by these multivariate 
analyses to predict graduate enrollment are presented below. As before, the coeffi- 
cients are standardized; the sample consists of all institutions. 

Biological Sciences: 

E = .29 b T - .23 b C + .14M a + .24 b O + .30 b F. 
Psychology: 

E = .76T b - .02C + .04M - .010 + .05F. 

In biology federal funds have a statistically significant effect on enrollment; this 
factor is more important than any other (although faculty size has almost as strong 
a coefficient). 10 

Federal funds do not appear to be significantly associated with enrollments in 
psychology. However, this finding may be the result of two factors in combination: 
the large number of masters candidates in psychology and that masters and Ph.D. 
candidates are lumped together in the available federal data. 



DEPARTMENT SIZE 

As noted earlier, department size was measured by the National Science Foun- 
dation manpower survey; we took the index of full-time scientists and engineers 
engaged in teaching. We then built a model in which the following factors were 
hypothesized to determine department size: 

• Type of control — i.e., public or private (C); 

• The presence of a medical school (M); 

• Federal funds to the discipline (F); 

• Other outside sources of support of the discipline (0); 

The multivariate equation relating these structural and financial variables to 
the dependent variable, size of department teaching staff (T), is given below: 



T t = B L C + B 2 M + B 3 t _ J + B 4 F t _ k . 



Tables 7 and 8 present information indicating the degree to which this model 
explained the variation in teaching faculty for the biological sciences and psychol- 
ogy. The R 2 statistics are lower than those for the previous two departmental in- 
dices. Note that our model as summarized in Fig. 1 postulated fewer determinants 
for faculty size than for the student variables. 

In biology the model fits better for universities without medical schools than for 
those with medical schools. In psychology the model does somewhat better in institu- 
tions that are private and those that have medical schools. 

The analysis of data for all institutions in 1975 yielded the following regression 
equations for the two fields: 

10 The funding coefficient reported here indicates only the direct effect of funding upon enrollments, 
given a certain faculty size. The indirect effect of funding on enrollment through its effect on faculty size 
is not reflected in this funding coefficient. Consequently, the net effect, both direct and indirect, is larger 
than that indicated by the coefficient. 



18 



Table 7 

PREDICTION OF TEACHING MANPOWER IN THE BIOLOGICAL SCIENCES 
„2 



(R 



Percentage of Variance Explained) 



1969 1970 1971 1972 1973 1974 



All Institutions 



148 



.30 



.41 



Table 8 



PREDICTION OF TEACHING MANPOWER IN PSYCHOLOGY 



(R 



Percentage of Variance Explained) 



.31 



.37 



Public 


93 


.32 


- 


.54 


- 


.11 


.35 


Private 


55 


.32 


- 


.40 


- 


.46 


.41 


With Medical School 


68 


.16 


_ 


.29 


_ 


.21 


.28 


Without Medical School 


80 


.46 


- 


.51 


- 


.33 


.43 


Research Universities 1 


49 


.08 


_ 


.19 


- 


.17 


.19 


Research Universities 2 


40 


.47 


- 


.39 


- 


.47 


.42 


Doctoral Granting 
















Universities 


47 


.42 


- 


.61 


- 


.43 


.36 







N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 




148 


.23 


- 


.27 


- 


.32 


.27 


Public 




93 


.16 


_ 


.26 


_ 


.31 


.21 


Private 




55 


.38 


- 


.33 


- 


.30 


.61 


With Medical School 




68 


.22 


_ 


.41 


- 


.46 


.37 


Without Medical School 




80 


.23 


- 


.20 


- 


.19 


.20 


Research Universities 


1 


49 


.21 


_ 


.29 


_ 


.38 


.30 


Research Universities 


2 


40 


.28 


- 


.38 


- 


.30 


.38 


Doctoral Granting 


















Universities 




47 


.14 


- 


.06 


~ 


.08 


.10 



Biological Sciences: 

T = -.14C + .26M b + .020 + .48 b F. 
Psychology: 

T = -.31C b + .10M - .040 + .41 b F. 

In each field the coefficient for federal funds is statistically significant at the .01 
level. Furthermore, in each case it is the most powerful determinant of faculty size. ' ' 

1 ' In the construction of our model, a particular problem was posed by the relationship between faculty 
size and enrollment. Those familiar with graduate education know that, to a certain degree, these two 
indices are interdependent; further, the relationship between the two is different in the public institu- 
tions than in the private. However, to postulate each as a determinant of the other would have created 
a model that was statistically and logically unresolvable. On the basis of previous research, the literature 
about graduate education, our field trips, and substantial accumulated knowledge about the functioning 



19 



While the NSF manpower data were the best available sources of longitudinal 
information about faculty size, some technical problems should be noted. Both post- 
doctoral and residents are included as well as faculty. We used that subset of 
full-time manpower in the department "primarily employed" in teaching. Certain 
obvious ambiguities exist in differentiating those faculty primarily engaged in 
teaching from those primarily engaged in research. In fact, some universities, espe- 
cially public institutions that are struggling with their legislatures, insist on report- 
ing all faculty in the teaching staff column. 

As a check on these problems we coded faculty size data from the American 
Council on Education quadrennial, American Universities and Colleges. Here, too, 
there were some technical problems: faculty size data were available for only two 
academic years — 1966-67 and 1970-71 — and the categorization of departments from 
school to school was ambiguous and inconsistent. We then repeated those biological 
science analyses in which faculty size was predicted or was used as a predictor, 
substituting the ACE measure. We found we were able to predict this new faculty 
size measure at about the same level as before and that group comparisons of the 
relationship between manpower and federal funding remained the same. Similarly, 
our ability to predict Ph.D. production remained the same when we used the new 
faculty size variable; our ability to predict graduate enrollment improved very 
slightly. 



COMPARISON OF DIFFERENT TYPES OF INSTITUTIONS 

In the sections above we assessed the adequacy of our model in explaining the 
variation in each of three indices of departmental structure and function. In addi- 
tion we determined whether federal funding was a significant factor in estimating 
each criterion and compared its predictive power with that of the other variables 
in the model. A related question is how the effects of federal funding vary across 
different sectors of higher education. In Tables 9-14 unstandardized regression coeffi- 
cients representing the relationship between (a million dollars of) federal funding 
and each of the three indices in each field are presented for the following subgroups 
within higher education: 

• Public or private; 

• Universities with medical schools and those without medical schools; 

• Each of the three leading Carnegie Commission categories. 

By using these subsamples we hoped to examine interaction effects and thus to 
trace the unique configurations associated with type of control and presence of a 

of academic life sciences departments, we postulated faculty size as a determinant of enrollment, not vice 
versa. 

In this chicken-egg problem, it is clear that federal funding nourishes both. To the degree that prior 
enrollment (for example, in public institutions) is a determinant of faculty size, the equations reported 
in this section may somewhat overestimate the federal funding coefficient. Conversely, had faculty size 
not been postulated as a determinant of enrollment it is likely that the coefficients for predicting 
enrollment on the basis of funding would have been larger. However, examination of our results indicated 
that this factor would not have changed the direction of the comparisons we report between public and 
private institutions. 



20 

Table 9 

RELATIONSHIP OF FEDERAL FUNDING TO PH.D. PRODUCTION IN THE BIOLOGICAL SCIENCES 
(Unstandardlzed Regression Coefficients) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 


148 


2.07 b 


.81 


2.04 b 


1.20 b 


1.18" 


.98 b 


Public 


93 


• 79 h 
2.64 b 


• 42 h 
1.87 b 


1.55 

3.22 


• 85 h 
2.51 b 


1.65 b 
.77 b 


1.26" 


Private 


55 


.98 a 


With Medical School 


68 


2.36 b 
2.88 


.81 


2.37 b 


• 74 h 
2.50 b 


• 91 b 
1.94 b 


.91 a 


Without Medical School 


80 


1.12 


1.15 


-.42 


Research Universities 1 


49 


2.31 b 
3.46 b 


.33 


1.87 3 


1.13 


1.43 b 


.92 


Research Universities 2 


40 


3.18 


2.55 


-1.52 


.00 


2.29 


Doctoral Granting 
















Universities 


47 


1.37 


3.35 


.76 


3.10 


1.44 


.00 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



Table 10 



RELATIONSHIP OF FEDERAL FUNDING TO ENROLLMENT IN THE BIOLOGICAL SCIENCES 
(Unstandardized Regression Coefficients) 





N 


1969 


1970 


1971 


1972 


1973 


All Institutions 


148 


26.72 b 


30.10 b 


25.67 b 


17.53" 


13.08 b 


Public 


93 


34.65" 


42.30 b 


31.93 b 
16.34 b 


18.42 b 
15.98 


14.13 b 


Private 


55 


14.04 a 


15.23 a 


8.90 a 


With Medical School 


68 


24.11 b 


29.09 b 
29.87 


24.57 b 
25.27 


13.60 3 
25.76 b 


9.28 u 
26.94 b 


Without Medical School 


80 


27.83 a 


Research Universities 1 


49 


19.30 v 
35.66 b 


21.93 a 


18.41 


9.85 


9.20 


Research Universities 2 


40 


30.97 a 


18.34 


18.26 


6.69 


Doctoral Granting 














Universities 


47 


-15.27 


13.61 


-16.73 


-17.92 


-19.92 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



Table 11 



RELATIONSHIP OF FEDERAL FUNDING TO TEACHING MANPOWER IN THE BIOLOGICAL SCIENCES 
(Unstandardized Regression Coefficients) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 


148 


12.10 b 


- 


13.03 b 


- 


6.83 b 


8.20 b 


Public 


93 


9.52 b 
13.37 b 


_ 


15.76 b 


_ 


3.24 u 
12.11 b 


6.46 b 
9.33 


Private 


55 


- 


9.49 b 


- 


With Medical School 


68 


11.57 b 
15.06 b 


_ 


12.61 b 
14.61 b 


_ 


6.98f 
6.38 b 


9.70 b 
6.40 b 


Without Medical School 


80 


- 


- 


Research Universities 1 


49 


7.37 


_ 


11.40 b 


_ 


6.24 v 
18.10 b 


7.57 a 
12.79 b 


Research Universities 2 


40 


7.63 


- 


12.30 3 


- 


Doctoral Granting 








29.23 b 








Universities 


47 


1.53 


- 


- 


-4.48 


1.50 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



21 



Table 12 

RELATIONSHIP OF FEDERAL FUNDING TO PH.D. PRODUCTION IN PSYCHOLOGY 
(Unstandardized Regression Coefficients) 







N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 




148 


1.72 


- .42 


5.53 b 


1.45 


4.14 a 


6.72 b 


Public 




93 


.41 


.00 


7.18 3 


2.05 


5.27 b 


7.43 b 


Private 




55 


7.07 3 


- .71 


3.59 


.73 


2.10 


-5.91 


With Medical School 




68 


4.12 


- .75 


4.21 
11.89 


2.13 


3.49 


7.49 b 


Without Medical Schoo] 




80 


2.95 a 


1.89 


3.33 


4.74 


5.23 a 


Research Universities 


1 


49 


4.24 b 


-2.89 


.50 


-1.39 


.00 


3.67 


Research Universities 


2 


40 


1.24 


- .86 


5.99 


.00 


5.72 


6.47 


Doctoral Granting 


















Universities 




47 


.00 


-4.96 


.00 


4.14 


6.98 


2.77 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



Table 13 

RELATIONSHIP OF FEDERAL FUNDING TO ENROLLMENT IN PSYCHOLOGY 
(Unstandardized Regression Coefficients) 





N 


1969 


1970 


1971 


1972 


1973 


All Institutions 


148 


- 24.07 


-11.79 


- 41.30 3 


- 5.97 


8.54 


Public 


93 


- 35.81 


- 7.60 
-86.93 


- 56.16 a 


-15.45 


3.78 


Private 


55 


8.38 


- 34.82 


- 7.51 


.00 


With Medical School 


68 


- 36.27 


- 7.49 


-103. 46 b 


-11.89 


8.72 


Without Medical School 


80 


- 9.30 


- 9.49 


- 10.86 


.00 


16.06 


Research Universities 1 


49 


- 14.52 


3.28 


.00 


- 2.14 


.00 


Research Universities 2 


40 


- 30.19 


-11.82 


-180. 72 a 


-25.13 


30.92 


Doctoral Granting 














Universities 


47 


-148.89 


-57.27 


6.30 


- 7.84 


43.44 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



Table 14 

RELATIONSHIP OF FEDERAL FUNDING TO TEACHING MANPOWER IN PSYCHOLOGY 
(Unstandardized Regression Coefficients) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


All Institutions 


148 


16.44 b 


- 


22.15 b 


- 


18.57 b 


15.31 b 


Public 


93 


17.02 b 
16.65 b 


_ 


29.35 b 


_ 


15.62 b 
22.47 


10.89 a 
25.09 


Private 


55 


- 


9.10 


- 


With Medical School 


68 


18.41 b 
16.80 


_ 


16.11 b 
32.40 b 


_ 


14.79 b 
22.76 


9.49 a 


Without Medical School 


80 


- 


- 


14.30 


Research Universities 1 


49 


18.62 a 


_ 


14.50 


_ 


14.47 b 


9.44 


Research Universities 2 


40 


9.15 


- 


25.97 a 


- 


29.53 a 


28.55 


Doctoral Granting 
















Universities 


47 


2.40 


— 


-9.62 


- 


17.80 


10.11 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



22 



medical school (beyond the effects reflected in their coefficients when they were 
represented above as dummy dichotomous variables above). 12 

We included the Carnegie groupings to test some hypotheses that have been 
proposed about the distribution of scientific resources among the universities in this 
country. The concentration of federal support in a limited set of institutions that 
tend to be clustered geographically has been a persistent political issue for at least 
15 years. A major 1960 report of the President's Science Advisory Committee, known 
as the Seaborg Committee (after Glenn T. Seaborg, Chancellor of the University of 
California, Berkeley, and chairman of the committee) recommended, "over the next 
15 years the United States should seek to double the number of universities doing 
generally excellent work in basic research and graduate education. 13 Among other 
things, concern about this issue led to the creation of the NSF Science Development 
Program of the 1960s, sometimes referred to as the "Centers of Excellence" grants. 
This program flourished during President Johnson's administration and was consis- 
tent with his philosophy of geographic diffusion of funds. In fact a 1965 executive 
order on that subject was closely tied to the dispersal of NSF funds for Science 
Development. During the more recent period of retrenchment in graduate educa- 
tion, one of the national debates has been whether limited resources should be 
distributed widely or reserved for only elite institutions. 

In the wake of the crisis that hit graduate education, it frequently was asserted 
that the elite leading universities were "responsibly" reducing their graduate en- 
rollments and Ph.D. production while "lesser" institutions continued to confer doc- 
torates at a lively rate. This particular theory stated that the result would be a 
reduction in the quality of the Ph.D.s produced nationally, the implicit assumption 
being that better quality Ph.D.s are generated by the leading institutions. 14 Perhaps 
the most eloquent expression of this fear was contained in the highly publicized 
Newman report on graduate education. 15 All of these assertions are subject to 
empirical test and have not been examined previously for the life sciences. It seemed 
to us that the key test of these hypotheses required use of the multivariate conceptu- 
al model we have elaborated here. That is, the central question is not the number 

12 In the equations for public institutions and for private institutions the dichotomy about medical 
schools was retained; similarly, in the equation for the two medical school groups the dichotomy about 
type of control was included. Both dichotomies were included in the equations for the Carnegie groupings. 

13 President's Science Advisory Committee, Scientific Progress, the Universities, and the Federal 
Government, The White House, Washington, D.C., November 1960, p. 28. 

14 The most widely used assessments of graduate departments were two studies conducted by the 
American Council on Education. (A. M. Cartter, An Assessment of Quality in Graduate Education, 
American Council on Education, Washington, DC, 1966; and K. D. Roose and C. J. Andersen, A Rating 
of Graduate Programs, American Council on Education, Washington, D.C., 1970.) These evaluations, 
based on peer ratings of the quality of graduate faculty, not only have been widely used to establish an 
academic pecking order but also have become points of reference among federal officials, university 
administrators, and scientists to infer growth and change in the capabilities of specific science depart- 
ments. 

Several investigators have addressed themselves to discovering the objective correlates of the (subjec- 
tive) ACE ratings. David Drew and Ronald Karpf ("Evaluating Science Departments: A New Index," The 
Rand Corporation, P-5521, Santa Monica, California, October 1975) tested a number of objective indices 
and found that rate of publication in key journals predicted the ACE quality rating almost perfectly — i.e., 
with a correlation of .91. Their findings confirm the results of some prior investigations that indicated 
the ACE rankings favored larger departments, a failing that would be corrected by the use of per-person 
indicators. In addition, they note the great need for an effective means of assessing the quality of teaching 
in a department. 

15 U.S. Department of Health, Education, and Welfare, Report on Higher Education: The Federal 
Role — Graduate Education, Frank Newman, Chairman, Washington, D.C., 1973. 



23 



of Ph.D.s or the enrollments ofleading and less outstanding institutions, but rather 
how sensitive these indices are to federal funds in both categories of schools. Conse- 
quently, we conducted these analyses to test the effects of federal funds upon doctor- 
ate production, graduate enrollment, and manpower — i.e., teaching staff— in the 
three categories of institutions that constituted the bulk — i.e., all but thirteen uni- 
versities — of our sample: 

• Those institutions listed as "Research Universities 1" by the Carnegie 
Commission on Higher Education 

• Those universities listed as "Research Universities 2" 

• Those universities listed as "Doctoral Granting Universities 1" 

Of the remaining 13 institutions, nine had been selected for the full sample because 
they had medical schools; four others were parts of multi-campus universities. 

Numbers within a column in Tables 9-14 are comparable such that the criterion 
can be compared directly across sectors of institutions. The superscript provides a 
preliminary indication as to whether the coefficient represents a statistically signifi- 
cant relationship. The significant coefficients then can be read as reflecting the 
increment in, for example, Ph.D. production uniquely associated with a million 
dollars additional federal funding, independent of all other variables in the model. 

Examination of these tables reveals a mixed profile with respect to type of 
institutional control. Graduate enrollments in biology consistently were more close- 
ly tied to federal funds in the public than in the private sector (Table 10). In addition, 
the association of funding with enrollments declined over time in both sectors, but 
the drop occurred sooner in the state institutions. Thus, the time of responsiveness 
to funding changes may vary between the public and the private domains. 

Ph.D. production in both fields suggested a stronger tie to federal funds in the 
private sector during the early years and in the public sector during the later years. 
(Tables 9 and 12). 16 

In the more recent years, faculty size has been more associated with federal 
funds in private institutions that at the state schools (Tables 11 and 14). 

The presence or absence of a medical school had no consistent relationship to 
the effect of federal funds. However, as noted earlier, our analysis of this factor was 
greatly hampered by limitations in the available federal data. It was impossible to 
isolate, for example, graduate enrollments in a basic science at the main campus 
from those at the medical school. 

The direct effects of federal funds did not vary substantially between leading 
research universities and doctoral granting institutions. Furthermore, funding, en- 
rollment and Ph.D. trend data on biology and psychology revealed very similar 
patterns for the leading and lower-ranked institutions. This analysis, of course, is 
not a test of whether leading and lesser universities respond differently to job market 
trends but is limited to the direct effects of federal funding. 

While formal control was included as a control variable, sample size limitations 
precluded analysis of funding effects differentially by type of control and level of 
institution. The few leading-lesser differences evident in these tables may reflect 
public-private variation at the two levels of quality. 

16 One reviewer of this report, Donald Stewart, commented that this finding jibes with an observation 
made in a research project being conducted by Martin Myerson and colleagues that the past decade has 
witnessed the overtaking of the private universities by the public institutions. These structural changes 
in higher education may be reflected in the statistics reported here. Martin Meyerson et al., The Future 
of Research Universities, forthcoming 



24 



CUMULATIVE EFFECTS OF FEDERAL FUNDING 

The analyses reported above focused on the direct effects of federal funding upon 
Ph.D. production and enrollment. In addition, as indicated in Fig. 1 above, funding 
has an indirect effect on each through its effect on faculty size and has additional 
indirect effects on Ph.D. production through enrollment. In our next analysis, the 
combined direct plus indirect effects of federal funding on Ph.D. production and 
enrollment were calculated for the most recent year. These results provide the best 
estimate from these data of the current cumulative federal effect on each of these 
variables. (Note that the model postulates no indirect effects of federal funding on 
faculty size through mediating variables.) Table 15 presents those results for the 
biological sciences. Of course, the cumulative effect of federal funding exceeds the 
direct effect. Note that the direction of the differences between types of institutions 
— e.g., public or private — remains unchanged. 

Perhaps the most interesting finding in the table are for the three Carnegie 
categories. Enrollment levels in leading institutions are not tied to federal funds. No 
difference among the Carnegie levels is found in the link between enrollment and 
funding. However, the group of institutions at which federal funding is most closely 
tied to Ph.D. production may not be the institutions represented by the Carnegie 
categories "Research Universities 1" and "Doctoral Granting Institutions 1." They 
may be those universities in the intermediate "Research Universities 2" category. 



Table 15 

CUMULATIVE FEDERAL FUNDING EFFECTS IN THE BIOLOGICAL SCIENCES 
(Unstandardized Regression Coefficients) 







Graduate 


Ph.D. 






Enrollment 


Production 




N 


1973 


1974 


All Institutions 


148 


18.07 b 


2.78 b 


Public 


93 


23. 63 b 
11.74 b 


3.63 b 
2.12 b 


Private 


55 


With Medical School 


68 


12.94 b 
36.29 


2.20 b 
4.71 b 


Without Medical School 


80 


Research Universities 1 


49 


10.75 


1.94* 
4.58 b 


Research Universities 2 


40 


16.58 


Doctoral Granting 








Universities 


47 


-6.39 


-.17 



Coefficient significant at .05 level. 
Coefficient significant at .01 level. 



RELATIONSHIPS BETWEEN FEDERAL SUPPORT AND 
OTHER SOURCES OF DEPARTMENTAL INCOME 

In the above discussion, we included both federal support of each discipline and 
nonfederal outside support. The latter category comprises both private and state 



25 



funds for research. Additionally, we retrieved data from the NSF expenditure sur- 
vey on "departmental research and instruction," the internal university funds for 
the discipline. The NSF data were available only for the aggregate category "life 
sciences" — which contains biology, agriculture, clinical medical fields, and a miscel- 
laneous category — and psychology; consequently, the statistics reported in this sec- 
tion are for those two discipline categories. 

A final financial datum was the total flow of NIH research money to the disci- 
pline as retrieved from NIH's "IMPAC" file on awards. 

We wanted to study the relationships among these various sources of depart- 
mental income over time. 17 Do other nonfederal outside funds for research follow 
federal funds? Is the pattern of NIH support over time parallel to that for all federal 
support? 18 

The trends in each of these financial indices over the past decade are presented 
for each category of institution in Tables 16 and 17 (life sciences) and Tables 18 and 
19 (psychology). 19 Examination of these tables reveals considerable information 
about the flow of dollars to the academic research and instruction efforts during the 
past decade. 

Federal and nonfederal funds appear to vary in parallel patterns, a phenomenon 
that could result from a number of causes. However, these simple trend data do not 
support the notion that one federal dollar is seed money for the attraction of several 
nonfederal dollars. 

A special analysis was conducted to test the degree to which an increase in 
federal funds in a given year is tied to an increase in nonfederal funds in a subse- 
quent year. Specifically, the change in federal funds to the biological sciences be- 
tween 1972 and 1973 was correlated with the change in nonfederal funds from 1973 
to 1974. The partial correlation (i.e., federal 1973 with nonfederal 1974 controlling 
for federal 1972 and nonfederal 1973) was not significant. This finding also does not 
support the idea that federal funds seed nonfederal. The partial correlation for 
public universities only also was not significant. Surprisingly, the partial correlation 
for private institutions was significant but negative. Although this is difficult to 
interpret, it may indicate a zero-sum situation in which a dollar of federal money 

17 Balderston has commented on some inherent problems in this type of investigation: "These funding 
agencies often wish to ensure that funds awarded are used for the purpose agreed, which is something 
that adequate financial stewardship and grant administration by the university can cope with up to a 
point. But, as several components of funding are used to support intertwined activities, funding agencies 
can never be quite certain that they are getting what they think they are for their money." F. E. 
Balderston, "Difficulties in Cost Analysis of Graduate Education," in National Board on Graduate Educa- 
tion, Federal Policy Alternatives Toward Graduate Education, Washington, D.C., January 1974, p. 96. 

18 Unfortunately, we did not receive these NIH data in time to incorporate them in our multivariate 
analyses; similarly, ADAMHA funding data arrived too late to be included in these trend analyses. 

19 Specific NIH figures should be compared with comparable statistics on federal and nonfederal 
funding in a given category and year only with extreme caution because the NIH statistics represent 
obligations, not expenditures, and reflect only the main campus, not associated health professional or 
medical schools. Finally, it was not possible to classify a number of NIH grants by department from the 
available descriptive information, and it was therefore necessary to omit significant amounts of NIH 
funding. In short, were it not for these factors, the actual NIH amounts would represent a much larger 
proportion of "all federal sources" for these disciplines. 

In computing the statistics in those tables, schools for which information from a given source was 
missing in a given year were omitted from that calculation only. An alternative approach, which we 
rejected in light of the very small number of missing observations, would have been to drop any school 
that had any item missing. This would have reduced the sample from 148 to 99 and severely biased our 
statistics. In fact, missing observations in these financial data were negligible except for internal funds 
in 1968 and 1970. A check of those means against the means that would have resulted from the alterna- 
tive approach showed little discrepancy. 



26 

Table 16 

TRENDS IN NIH SUPPORT TO UNIVERSITY LIFE SCIENCE DEPARTMENTS 
BY TYPE OF INSTITUTION * 

(Thousands of Dollars) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


1975 


All Institutions 


148 


380 


374 


425 


470 


456 


579 


643 


Public 


93 


323 


331 


363 


400 


388 


479 


538 


Private 


55 


476 


447 


532 


589 


572 


748 


819 


With Medical School 


68 


361 


355 


390 


421 


429 


544 


620 


Without Medical School 


80 


396 


390 


456 


513 


480 


609 


662 


Research Universities 1 


49 


864 


856 


994 


1078 


1064 


1344 


1531 


Research Universities 2 


40 


258 


256 


266 


303 


276 


344 


349 


Doctoral Granting 


















Universities 


47 


73 


64 


72 


94 


84 


113 


113 



*See text for discussion of these data. 



Table 17 



TRENDS IN FEDERAL AND OTHER SUPPORT TO UNIVERSITY 
LIFE SCIENCE DEPARTMENTS BY TYPE OF INSTITUTION 

(Thousands of Dollars) 



1964 1968 



1970 



1972 



1973 



1974 



All Institutions 148 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Public 93 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Private 55 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

With Medical School 68 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Without Medical School 80 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Research Universities 1 49 
All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Research Universities 2 40 
All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Doctoral Granting 
Universities 47 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 



2521 


3941 


4483 


4626 


5445 


5553 


1289 


1637 


2026 


2448 


2748 


3055 


- 


4042 


4359 


5318 


6100 


6883 


2089 


3135 


3814 


3732 


4427 


4463 


1603 


1948 


2367 


2660 


3093 


3488 


— 


4036 


4342 


5318 


6470 


7216 


3244 


5300 


5579 


6138 


7167 


7397 


763 


1115 


1467 


2089 


2164 


2322 


— 


4052 


4389 


5319 


5496 


6338 


4032 


6380 


7746 


8164 


9599 


9729 


1512 


1833 


2488 


3211 


3486 


3789 


" 


6166 


7015 


9303 


10772 


12053 


1255 


1788 


1681 


1619 


1915 


2004 


1101 


1465 


1628 


1799 


2121 


2431 


— 


2245 


2266 


1947 


2198 


2564 


5395 


8407 


10202 


10682 


12594 


12615 


2491 


3032 


4004 


4958 


5353 


5968 


— 


7329 


8883 


10309 


11757 


13138 


1761 


2769 


2653 


2576 


2962 


3197 


1197 


1637 


1858 


2118 


2500 


2797 


- 


3126 


3752 


3866 


4661 


4972 



564 


782 


821 


846 


1041 


1102 


352 


483 


538 


628 


816 


873 


- 


1548 


1455 


1818 


2090 


2581 



27 

Table 18 

TRENDS IN NIH SUPPORT TO UNIVERSITY PSYCHOLOGY DEPARTMENTS 
BY TYPE OF INSTITUTION * 

(Thousands of Dollars) 





N 


1969 


1970 


1971 


1972 


1973 


1974 


1975 


All Institutions 


148 


59 


59 


57 


63 


54 


82 


77 


Public 


93 


59 


54 


42 


48 


43 


70 


64 


Private 


55 


60 


68 


82 


88 


72 


101 


99 


With Medical School 


68 


91 


93 


82 


89 


69 


97 


103 


Without Medical School 


80 


33 


30 


35 


40 


40 


68 


55 


Research Universities 1 


49 


130 


133 


123 


134 


107 


160 


163 


Research Universities 2 


40 


39 


38 


39 


39 


43 


68 


66 


Doctoral Granting 


















Universities 


47 


18 


13 


16 


24 


20 


33 


15 



*See text for discussion of these data. 



Table 19 



TRENDS IN FEDERAL AND OTHER SUPPORT TO UNIVERSITY 
PSYCHOLOGY DEPARTMENTS BY TYPE OF INSTITUTION 

(Thousands of Dollars) 



1964 1968 



1970 



1972 1973 1974 



All Institutions 148 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Public 93 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Private 55 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

With Medical School 68 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Without Medical School 80 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Research Universities 1 49 
All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Research Universities 2 40 
All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 

Doctoral Granting 
Universities 47 

All Federal Sources 
Nonfederal Outside Sources 
Internal Funds 



141 


269 


282 


305 


326 


322 


23 


61 


70 


87 


89 


93 


~ 


423 


520 


653 


674 


756 


129 


286 


310 


323 


345 


346 


23 


72 


86 


101 


103 


142 


_ 


466 


577 


772 


791 


884 


162 


239 


236 


274 


293 


281 


24 


42 


44 


63 


65 


62 


— 


335 


416 


461 


481 


541 


213 


377 


405 


425 


455 


454 


30 


71 


99 


113 


103 


111 


— 


481 


611 


827 


830 


971 


82 


174 


176 


203 


216 


209 


17 


52 


45 


65 


76 


78 


— 


371 


446 


507 


540 


573 


297 


539 


590 


636 


682 


669 


33 


108 


157 


195 


205 


214 


- 


651 


830 


1071 


1024 


1152 


115 


229 


206 


225 


242 


240 


38 


51 


42 


57 


43 


53 


- 


382 


519 


543 


598 


724 



1 


77 


88 


97 


104 


103 


6 


33 


19 


19 


25 


18 




257 


312 


412 


447 


457 



28 



substitutes for nonfederal funds. (Bear in mind that nonfederal outside research 
support includes only private sources for the private institutions, but includes pri- 
vate and state research support in the public institutions.) While a definitive inter- 
pretation of this finding was not possible in the limited time available, the possible 
explanations provide fascinating hypotheses for future research. 

A full examination requires exploring the effects of the nonfederal money that 
presumably is drawn by the magnet of federal dollars. The analyses reported earlier 
in this study consistently indicated that nonfederal outside research funds have a 
weaker tie to the basic indices of departmental structure and function than do 
federal funds. 20 

20 One exception in biology: In 1969, 1972, and 1973 the effects of nonfederal funding exceeded those 
of federal funding in the "Doctoral Granting Universities," perhaps reflecting the influence of state funds 
at public institutions. 



IV. FIELD TRIP RESULTS 



The field trips were conducted for three main reasons. First, the visits aided in 
interpreting the results of our data analysis. Second, we wanted to explore some of 
the causal relationships between shifts in federal research funding policy and 
changes in universities that we could not examine in our data analysis. Third, and 
most important, we wanted to determine the variety of institutional factors in 
universities that mediate the effects of federal research funds, and to find out how 
strong these factors are in relation to federal policy. 

Because we visited only ten universities and could spend only a limited time at 
each one, the results from our field work are less well-grounded than those from our 
data analysis. Therefore, they are not clearly generalizable to the entire university 
community affected by federal research funds. Our findings should be regarded as 
providing a background against which to assess the results of the data analysis and 
some suggestive insights into the workings of federal research funds on universities, 
but not as substantial conclusions. However, since our sample of universities was 
selected to be representative of the research university community, we are confident 
that our results have some validity and, at a minimum, suggest topics that merit 
additional research. 



UNIVERSITY ADMINISTRATION 

As noted above, the universities we visited represented a range of administrative 
structures. Five were public universities receiving substantial funds for instruction 
from state government, and five were private universities with most of their funds 
for instruction coming from private sources. Five of the ten universities had a 
medical school and five did not. We chose not to include university size as an explicit 
criterion in selecting our sample. However, by including the total amount of federal 
research funds received in a given year as a selection criterion, we had universities 
that spanned the full range of sizes. One university had a total student enrollment 
under 2000 while at the largest institution over 40,000 students were enrolled. 

As expected, the formal organizational structures of the larger universities were 
more complex, primarily in that they had an additional layer of vice-presidents (for 
academic affairs, health affairs, sponsored research, and so forth) over the usual 
corps of deans (for faculty, arts and sciences, the graduate school, and so forth). In 
addition, the larger universities tended to have more, and larger, specialized support 
offices of various kinds. 

A notable (though difficult to describe) difference among the universities we 
visited was the strength of the central administration. This bore little relationship 
to a university's size or complexity, or simple measures of authority. In our view, 
strong central administrations were both forward-looking and balanced with regard 
to the development of departments and had specific knowledge at their fingertips 
concerning their institution's budget, faculty size, and so forth. In strong administra- 
tions power was shared among several top-level administrators in close communica- 
tion with each other; they were more skilled at dealing with department chairmen 

29 



30 



and faculty members and at changing policy without generating excessive conflict. 
A strong central administration was a recent development in several of the universi- 
ties that we visited, having been caused by the onset of financial problems that led 
to the appointment of new leadership. 

There did not appear to be any correlation between either the organizational 
structure or the strength of a university's central administration and the quality of 
its departments of life and behavioral sciences. High quality life science depart- 
ments were seen in large, complex universities and in smaller universities with 
leaner and less formal administrations. 

Larger universities tended to have larger sponsored research or grants and 
contracts offices to perform more support functions. Traditionally, the primary 
responsibility of this office was keeping the financial records needed by the univer- 
sity or required by the federal government; another related and important function 
has been tracking the university's indirect cost expenses and negotiating a recovery 
rate with the federal government. 

In several of the universities we visited, the quality of the financial records kept 
by sponsored research offices has been spotty in the past, but partly as a result of 
increasing university and federal demands for improved financial information, 
these offices are developing more sophisticated accounting systems. 

A new role being assumed by sponsored research offices is assisting faculty in 
the preparation of research proposals through providing editorial, graphic, and 
clerical services, and in the location of potential funding sources. Some offices have 
developed to the point where selected staff members are assigned responsibility for 
keeping track of the research needs of certain federal agencies and notifying faculty 
of opportunities. Some have gone a step further and become involved in initiating 
and coordinating the development of large-scale projects, including recruiting facul- 
ty, providing start-up resources, and managing the submission of proposals for 
federal funding. These new roles of sponsored research offices can be seen as re- 
sponses on the part of universities to the shift in federal funding toward contract 
research, and the increasing complexity and intensity of competition for federal 
research funds. In general, our visits suggest that sponsored research offices tend to 
be more active in the area of contract research than in investigator-initiated re- 
search and to have developed the furthest in large, public universities receiving 
substantial federal funds, not necessarily those that are the most prestigious. 

A clear picture of the extent to which these offices increase a university's success 
in obtaining federal research funds did not emerge from our field trips. Administra- 
tors tend to see sponsored research offices as indispensable to obtaining research 
support in an increasingly complex environment, while faculty tend to see them as 
being of little value other than in providing assistance in the preparation of propos- 
als. There was no clear relationship between the size and strength of the sponsored 
research office and the quality of a university's life science departments. 

The directors of four of these offices were university vice-presidents (typically 
vice-president for research); three of them appeared to be highly influential in 
developing research programs in their universities. Again, these were not necessari- 
ly the largest research offices. The influence of these individuals was not so much 
their ability to find federal funds as their ability to work with faculty and organize 
interdepartmental research activities. Often this involved securing federal research 
support after a team of professors had been organized to conduct research on a topic. 



31 



Frequently university funds were used to get these teams started. By contrast, the 
other directors of sponsored research were oriented largely to tracking federal 
funding policy and standard administrative functions; they appeared to have much 
less effect on research in their university. In short, the research office directors who 
had high status in their universities and were internally oriented to developing the 
substance of their university's research programs appeared to be having more effect 
than directors oriented largely to dealing with the government. 

Our observations have been that the administrative structure of a university is 
unrelated to the quality of its life science departments, except that strong support 
from the central administration appears necessary for departmental development 
and that a vice-president for research can perform a useful research program devel- 
opment function. In other words, departmental quality probably depends far more 
strongly on factors other than the structure of the central administration. 



INDIRECT COSTS 

There was great institutional variation in the stress or lack of it felt by the 
central administration with respect to recovery of indirect costs from the federal 
government. Two public institutions illustrate the extremes. At one, very little 
federal support is received and indirect cost is not an issue. In addition, at that 
institution, the state routinely makes up the difference between what the govern- 
ment pays in indirect costs and what a full recovery from the point of view of the 
school would be. As a result, this university is less motivated to try to extract every 
last dollar from the federal government. At the other extreme is a university that 
does a huge volume of federal research where the administrators feel that they are 
losing millions of dollars annually by not recovering their full indirect costs. Legisla- 
tors at the state capitol are acutely aware that the state is picking up the difference 
and want to know whether the university can justify doing this much research, 
making it even more difficult for scientists at this leading institution to propose 
doing new research with federal support. 

Indirect cost is the cause of lively controversy between the academic community 
and the federal government and within universities. The latter controversy centers 
about the distribution within the institution of those costs recovered from the fed- 
eral government. Scientists believe that they earn all indirect costs associated with 
their grants and contracts; administrators tend to feel the funds belong in a general 
pool and cover a multitude of institutional costs not directly associated with a given 
researcher. There is extreme variation in policy from institution to institution. At 
some the funds are turned back directly to the investigator. At some they are kept 
by the central administrator and spent in whatever manner the leaders of the 
institution feel is necessary. At some a percentage is kept and the remainder re- 
turned. To a degree, the manner in which indirect costs are handled is a function 
of the network of power relationships within the institution. For example, at one 
leading research university all indirect costs are kept by the central administration 
as discretionary funds — except for one powerful, fairly independent unit, which 
keeps all of its own indirect costs. 

Another problem mentioned in several universities is the difficulty caused by the 
different approaches to indirect costs taken by the National Science Foundation and 



32 



the National Institutes of Health. 1 NIH awards grants in terms of direct costs. NSF 
often awards an amount for the total budget. If this is less than that originally 
proposed by the investigator, the breakdown of this new, smaller total budget figure 
into direct and indirect costs is a matter left to be negotiated between the institution 
and the faculty member. This, obviously, often causes stress between them. 

Weaker research offices tend to get caught between their faculty and agencies 
that negotiate proposed budgets with principal investigators on the basis of total 
rather than direct costs, so that cutbacks required by the agency often come dispro- 
portionately from indirect costs. Strong sponsored research offices have standard 
policies and are able to enforce them, so that there is little difference in the actual 
rate of indirect cost recovery across agencies. 



FINANCIAL CONDITION OF UNIVERSITIES 

All of the universities we visited are having financial problems to differing 
degrees. Two of the private institutions had deficits of several million dollars in their 
current operating budgets and were in serious trouble but hoped to regain financial 
health through proposed adjustments. The budgets of the three other private univer- 
sities were either in the black or sufficiently close to being balanced that small 
adjustments would eliminate any deficits. The five public universities had all been 
through a decade of rapid growth in revenues and student enrollment but were in 
various stages of having to cope with smaller increases in their total revenues. The 
prospect was that for the foreseeable future economic pressures on state revenue 
would preclude any substantial increases in funds for higher education. 

Our impression based on these field trips is that the "new depression," which 
hit graduate education at the beginning of this decade, was experienced first by the 
private institutions. The public universities were cushioned from the shock some- 
what by continuing increases in enrollments that states were willing to fund; only 
recently, with the growing disenchantment of some state legislators, are they an- 
ticipating the same kind of financial strains. Perhaps the most dramatic example 
we saw was a state institution that grew at an astonishing pace in the sciences 
during the early 1970s, when many private institutions were beginning to experi- 
ence financial difficulties. Throughout this period, both this state's population and 
enrollments in the university grew rapidly, and there was strong support in the state 
capital for increasing the university's budget to pay for these increases. But opposi- 
tion to further large increases in the university's budget emerged recently in the 
legislature, and relations are likely to be even more difficult in the future. An 
indication of this change is that the state legislators refused to appropriate money 
for shelves and other basic equipment to furnish a library that they had generously 
appropriated $10,000,000 to construct several years earlier. As a result, the library 
stands completed, but empty and unused. 

Since all universities receive funds from a variety of public and private sources 
and spend in a broad range of categories, it is hard to generalize about the causes 
of current financial problems. With some exceptions, all universities that we visited 

1 For a detailed examination of differences among federal agencies in the administration of research 
and development, see John Wirt et al., R&D Management: Methods Used by Federal Agencies, D. C. 
Heath, Lexington, Mass., 1975. 



33 



have been experiencing increases in most sources of revenue. However, there are 
strains because these sources of revenue are not increasining at the same rates as 
in the past. The most pervasive cause of financial problems was inflation, which was 
obviously occurring in all expenditure categories. Salaries and wages, which are the 
bulk of university expenditures, are up substantially; utility bills have doubled and 
tripled; construction costs are up; and the same piece of research equipment costs 
far more today than it did in the past. For example, the chairman of the biology 
department at a private institution noted that the cost of an electron microscope had 
more than doubled in the past five years; he added that only 15 to 20 percent of the 
increase resulted from increased sophistication in the instrument. 

Universities with strong central administrations appeared to be coping more 
effectively with their financial problems; that is, their budgets were closer to being 
balanced and instead of being buffeted by successive financial crises, they have 
moved aggressively but carefully in recent years to make the necessary cutbacks. 

Financial problems in the universities that we visited did not appear to be 
generally attributable to federal research funding. This would have been the case 
if institutions were experiencing substantial declines in their overall levels of feder- 
ally sponsored research, which then created problems in covering the salaries of 
personnel to whom the institution was committed. With one exception, total federal 
research funds received by the universities we visited had steadily increased over 
the last few years. Or, difficult readjustment problems could also occur if an institu- 
tion experienced severe cutbacks in funding for key large research facilities, even 
though overall levels of federal research funding have continued to increase. One 
of the universities in our sample had experienced such a decline in funding but was 
coping with the problem. 

Two indirect effects of federal research funding may be more serious. To the 
extent rates of reimbursement for the indirect costs of federally sponsored research 
are not sufficiently high to cover expenses, universities are making up the difference 
from their other sources of funds. Several of the universities we visited claimed that 
their indirect cost reimbursement rates were too low by a few percent, a substantial 
amount when total federal research support is several tens of millions of dollars. To 
the extent that science departments do not fully cover new research faculty with 
federal funds, inflation in their salaries represents a cost burden being borne by 
universities rather than by the government. 

At several institutions a special cash flow problem caused by federal government 
policies was cited as troublesome. There often is a great delay in the federal bureau- 
cracy between the time a grant is awarded and the time the paperwork has been 
cleared up and the money comes in. During this period the institution feels an 
obligation to allow the investigator to begin (or continue) his research. The institu- 
tion must temporarily advance the funds for this and loses interest on that money 
during the interim period. Enough funds fall into this category that the interest lost 
is a nontrivial amount at some universities. 

The financial situation at the departmental level of universities is clearly differ- 
ent and is directly related to federal research funding. In general, university depart- 
ments in the health-related sciences were continuing to obtain research funds and, 
as mentioned above, usually in increasing amounts. However, particularly in phys- 



34 

ics and most fields of chemistry the situation was bleak. 2 For example, in one 
university, which was otherwise highly research-oriented and experiencing a steady 
growth in federal funds, the chairman of the department of chemistry said that 70 
percent of his faculty who previously had federal grants were now unable to obtain 
support except in the area of analytic chemistry. The faculty members unable to 
obtain research funding were in what the department chairman called a "downward 
spiral": Without federal grants, they could not support graduate students and pur- 
chase equipment to conduct research; without conducting research, they cannot 
continue publishing in the journals; and, without publishing, it becomes even more 
difficult to obtain grants. 

An analogous though less protracted situation exists among the different fields 
of the life sciences in the universities that we visited. Funding for research is 
plentiful in the health-related areas of the biological sciences and psychology com- 
pared with other areas in these fields related, say, to the environment and agricul- 
ture. Many faculty that were interested in conducting research in these other areas 
were having great difficulty obtaining grants to support their work. Examples were 
limnology (fresh water biology) and photosynthesis. The National Science Founda- 
tion is the principal source of funds for those investigators. These faculty members 
have not found such agencies as the Environmental Protection Agency, the Depart- 
ment of Interior, and the Department of Agriculture to be as supportive. 



ORGANIZATION OF THE LIFE SCIENCES IN UNIVERSITIES 

Reflecting the breadth of the field of biological science and its direct relevance 
to such diverse fields as health, agriculture, and the environment, the patterns of 
organization of the biological sciences in university schools of arts and sciences were 
highly varied. One formal difference was whether the university had one depart- 
ment or division spanning all the traditional disciplines of the biological sciences 
(e.g., biochemistry, botany, and zoology) or separate departments. Biological science 
departments also existed in medical schools and in agricultural schools. These de- 
partments often duplicated the name if not the substance of academic activities in 
the university departments. 

Since the companion study by our colleagues at Rand focused on the effects of 
federal funding at medical schools, we did not visit medical school departments. Our 
work was concentrated on the life sciences at the main campus of the institutions 
we visited as well as at auxiliary, related health professional schools — e.g., the 
dental school, school of public health, agriculture school, veterinary school. How- 
ever, we did schedule some interviews at medical schools to explore how much 
activities there might have affected the relationship between federal funding and 
scientific research at the main campus. 

Four of the ten universities we visited were in various stages of consolidating 

2 The relative success of the life sciences perhaps should not be surprising. As far back as November 
1965, AlvinWienberg, a noted expert on relationships between the government and the scientific com- 
munity, hypothesized that the future volume of government expenditures for universities and research 
would probably hinge on three factors, one of which was expenditures on biological research. (The otner 
two were the discovery of ways to attack certain social problems and the role of the National Science 
Foundation.) Alvin Wienberg, "Government Allocations to Basic Research," in Harold Orlans (ed.), 
Science Policy and the University, The Brookings Institution, Washington, D.C., 1968, p. 159. 



35 



their separate biological science departments into a single department or division. 
Typically, before reorganization, one of the departments had had several faculty 
oriented to the traditional areas of biology — systematic biology, botany, zoology, and 
so forth. (Departments of psychology were not included in these reorganizations.) 
The primary theoretical reason for these efforts is that recent fundamental advances 
in understanding of basic biological processes (for example, the mechanisms of cell 
replication) have unified the traditional disciplines of biological science, requiring 
new forms of departmental structure. Typically, a university's objectives for reor- 
ganization were to shift faculty orientation from traditional to modern forms of 
biology, unify and modernize curriculum offerings, and increase funds for externally 
sponsored research. The bulk of these funds are available only in the new areas of 
inquiry. The rationale for reorganization as a way of achieving these objectives was 
to increase interaction among faculty in the new areas through breaking down 
traditional departmental barriers to communication, and to free resources for hiring 
new faculty trained in modern areas of biology by eliminating duplication in faculty 
positions, course offerings, degree programs, and research equipment and facilities. 

All four examples of reorganization in the universities we visited had been 
initiated by the central administration rather than the department chairmen or 
faculty. In fact, the resistance of the latter groups to this innovation has been great 
in all four cases . Although there are often some faculty who support reorganization, 
in general most department chairmen and faculty do not believe that the presumed 
benefits will accrue and fear that their specialties will not receive proper resource 
support in a single, large department. 

To date, these restructurings largely have failed in the universities we visited. 
In all four cases, it has either been a continuing process, involving a succession of 
reorganizations and department chairmen, faculty conflict, few new faculty hired, 
and no substantial increase in externally funded research; or the single department 
is a paper organization with little real influence on the university's biological science 
activities. 

The only universities we visited where many new life science faculty have been 
added and strong, high quality departments have been built over the last decade 
were universities with new departments. Previously, these universities had had few 
life science faculty. Two universities fell in this category and both had received 
institutional development awards from either the NSF or NIH. One school had 
received two such awards. The department chairmen in these universities believed 
that the institutional development awards had been instrumental in moving their 
departments forward. 

One of the universities where strength was developed in the life sciences has a 
single, large biology department (and a psychology department), and the other 
university has separate biological science departments (and no psychology depart- 
ment). In the former case, the single, large department functions well and is stable. 
The current department chairman believes that the unified structure facilitated the 
development of joint faculty research projects, led to an integrated educational 
program, and provided flexibility in utilizing space. Federally sponsored research 
has increased dramatically. This example indicates that the single department 
concept is workable. However, this department was built almost singlehanded by an 
extremely dynamic and able department chairman with strong support from the 
university administration. 



36 



Two other universities have strong, high quality, and modern life science depart- 
ments in their school of arts and sciences and always had them. Quality was main- 
tained over the years by strong departmental leadership, aggressive pursuit of 
federal research funds, and adherence to rigorous standards of appointments and 
promotions. Also, these universities have adhered to deliberate policies of concen- 
trating faculty expertise in selected areas and maintaining a steady flow of young 
assistant professors trained in emerging areas of the life sciences. 

The personality, drive, and leadership ability of a chairman can mean the 
difference between a department that grows toward excellence and one that lan- 
guishes. Our observation was that the importance of a dynamic chairman in develop- 
ing a given department greatly exceeds the effect of dynamic leadership in the 
central administration. 

It was striking that none of the four universities with strong life science depart- 
ments in their schools of arts and sciences had medical schools. Some of the universi- 
ties with medical schools had developed strong life science departments in the 
medical school, but not in the school of arts and sciences. Apparently, the presence 
of a medical school in these universities has inhibited the development of strong life 
science departments at the university's main campus. 3 In fact, two of the attempts 
by central administrations to reorganize and improve their life science departments 
were in institutions with medical schools. The main disadvantages of having strong 
life science departments only in the medical school are that such departments are 
strictly health related and are unlikely to provide service courses for students from 
other parts of the university. 

In summary, federal funds for life science research apparently have the greatest 
effect on departmental quality through providing the means for building new de- 
partments rather than through reorienting established ones. Furthermore, the de- 
velopment of strong departments is highly related to organizational factors within 
the university not directly susceptible to federal influence, such as the presence of 
a medical school, strong departmental leadership, and support from the central 
administration. 



FEDERAL FUNDING AND FACULTY 

In our field work we were interested in exploring how departments use federal 
funds to increase faculty size and quality, and the effects on faculty of the loss of 
funds. Many of the departments we visited had experienced substantial increases in 
federal research funds over the last decade, which made it possible to explore how 
departments use federal funds to grow. Only one university had experienced a 
substantial decline in total federal research funds in recent years. However, this was 
from a low funding level and its faculty are not in the top rank of life science 
departments. A few departments had experienced a slight decline and some a shift 
in the forms of funding, but these changes were not large enough to jeopardize 
numbers of faculty positions. In these departments a few tenured positions have 

3 One of our observations is that, regardless of departmental and institutional structure, first rate 
researchers will seek out other first rate researchers. For example, even at one public institution where 
a number of faculty and administrators commented on the lack of interaction between the medical school 
and the main campus, we heard of isolated examples of collaborative research between first rate scientists 
at the two places. 



37 



remained unfilled for longer periods of time than were normal in the 1960s, and 
promises by central administration for growth in tenured faculty positions made in 
the late 1960s have not materialized. (The shifts in forms of funding were primarily 
from training grants to research grants.) Thus, we could not directly observe the 
consequences for faculty of substantial declines in federal research funding. 

Departments under financial pressure, either from loss of federal funds or reduc- 
tions in university funding necessitated by general financial problems, tend to cut 
back first on nontenured faculty, as might be expected. But the cutbacks appeared 
to be more in the rate of promotion to tenure than in the total numbers of non- 
tenured faculty. For example, in one department that we visited, the number of 
assistant professors had not fallen substantially in recent years, but contracts for 
junior faculty have been changed from three years once renewable before a tenure 
decision, to one year renewable six times. This allows the department to be more 
selective in choosing tenured faculty and provides financial flexibility. We did not 
find any department that had not promoted any faculty to tenure in the last two 
years. 4 

Some of the life science departments that grew the most over the last decade 
built faculty size and quality directly with federal funds through agressively seeking 
larger amounts of federal grants and contracts, carrying a substantial proportion of 
faculty salaries on the funds received, and using the monies released to hire addi- 
tional and better faculty. Usually, these departments also had been successful at 
bargaining with their central administrations for increased internal university 
funds and for additional space, using as a rationale their demonstrated ability to win 
federal awards and improve themselves. But throughout the developmental period, 
these departments have continued to carry a much larger than normal proportion 
of their faculty salaries on federal funds. Whether the faculty carried on these 
government funds were tenured or nontenured varied by department. 

The number of departments that use such a "soft money" policy is small. Most 
faculty are firmly against having more than their summer salaries paid from grants 
and contracts, which they fear might disappear overnight, leaving them extremely 
vulnerable. There is also strong concern that long term overdependence on federal 
dollars is likely to erode the institutional independence fundamental to the concept 
of the university. 

Contrary to what might be expected, some of these departments in which faculty 
are strongly opposed to charging their salaries to federal grants grew as much over 
the last decade as departments that followed a soft money policy. They used their 
success in obtaining increased federal research funds to negotiate increases in facul- 
ty positions and other resources with their central administrations. 

Whether salaries are charged to federal grants and contracts depends more on 
faculty attitudes than on attitudes of the central administration. Often within the 
same university some departments are highly dependent on soft money and others 
are not. Departments in universities experiencing strong enrollment growth over 
the last decade had the best of both worlds, since increasing enrollments provided 
uncommitted funds that central administrators could use for rapidly converting soft 
money into hard money positions. In every case we saw, these were public universi- 
ties. 

4 One effect of the reduced rates of promotion to tenure is to make higher quality candidates available 
to the less prestigious departments. 



38 



We found no examples of life science departments following a risky, soft money 
policy and experiencing a substantial decline in federal funding and difficulty in 
covering faculty salaries. But the fact that federal support in the biological sciences 
has continued to grow is part of the explanation; we heard of physics and engineer- 
ing departments that have been highly dependent and have run into serious trouble. 
Still, most of those departments have continued to find sources of funding and 
maintain faculty size, though often with great difficulty. 

Another major use of federal research funds in life science departments is for 
purchasing research equipment. Neither private nor state universities have any- 
where near sufficient funds to permit the faculty to buy enough equipment to 
continue conducting research at the forefront of their disciplines. Faculty needs for 
equipment vary by research area, of course: a theoretical biologist may need little 
more than a pencil and paper; a biochemist may need a vast array of sophisticated 
equipment. However, most life science research is highly empirical, and faculty 
dependence on equipment is high. They would not stop doing research if no federal 
funds were available, but it might be far different in quality. 

Increasing technical sophistication and inflation have been driving the costs of 
research equipment up rapidly. There are few pieces of equipment (other than 
general supplies) that faculty can afford to purchase on a single project grant. 
Instead, money to buy equipment must be pieced together from several project 
grants and other sources. The faculty we interviewed report that the costs are 
increasing so rapidly as to create serious problems of acquiring and maintaining the 
stock of equipment they need for their research. This suggests that unless federal 
policy is changed to provide more funds for purchase and maintenance of equipment, 
the overall life science research effort may be increasingly hampered in the future. 

A departmental expense for which federal support cannot easily be used is 
providing startup resources for new assistant professors. The money available for 
this purpose varies greatly by departments and by universities. In some depart- 
ments, we interviewed assistant professors who received less than $500; in others, 
assistant professors received over $20,000 for this purpose. We were told about one 
young professor who had trouble getting hot and cold running water in her labora- 
tory. Some sponsored research offices provide funds, some graduate schools have 
general fund accounts and some universities have endowment income for this pur- 
pose. Occasionally, a department has been able to obtain a small foundation grant 
for a new professor. One had benefitted greatly from a small grant from the National 
Institute of Mental Health, the only example of direct federal support for startup 
we found. The more prestigious universities usually had substantially more funds 
for starting up young professors than the less prestigious and provided these funds 
more uniformly across departments. 

The ability of departments to attract young professors is probably highly depen- 
dent on their ability to offer startup funds, and, since the weaker departments have 
fewer of these funds, new federal programs to channel funds to lesser departments 
for startup expenses could be effective means of redistributing faculty quality among 
universities. NIMH's Small Grant program is a good example of such a program. 
Departments also could use funds they receive from NIH's General Research Sup- 
port program, which provides institutional support to departments based on the 
amount of NIH research funds received for this purpose, but our field trips suggested 
that universities tend to use these funds mostly for other purposes. 



39 



An important factor in explaining faculty size, unrelated to federal research 
support, is the department's undergraduate teaching load. In particular, biology 
departments have recently been experiencing significant increases in undergradu- 
ate enrollments and have used these increases as justification for obtaining addition- 
al faculty (and teaching assistantship) positions from their central administrations. 5 
By contrast, faculty size in biochemistry departments tends to be unrelated to 
undergraduate enrollments, since few offer any undergraduate courses. In short, 
department orientation can affect faculty size in a way that is independent of federal 
research funds. 

In summary, federal funds appear to affect faculty size through increasing de- 
partmental bargaining power with central administrations and making it possible 
to carry extra positions on outside funds. Federal funds affect quality through 
providing departments with the resources needed to purchase equipment. Less tan- 
gible, though probably more important, effects on departmental quality stem from 
the research activity itself. Faculty who conduct research stay in touch with 
progress in their disciplines and attract better graduate students and other faculty 
interested in research. How and how much departments use federal research funds 
to increase their size and improve quality are highly dependent on local departmen- 
tal norms and structural factors. 



FEDERAL FUNDING AND STUDENTS 

Another major use of federal funds in life science departments, as in most 
science departments, is financial support for graduate and postdoctoral students. 
Although we did not collect specific data, our overall impression was that almost all 
of these students receive financial support, the majority of them from federal funds. 
Departments generally prefer to support graduate students on training grants be- 
cause of the flexibility provided for students to choose their area of study. However, 
departments also support faculty and purchase equipment with training grant 
funds. 

As in most science areas, the commitment of faculty in biological science depart- 
ments to provide full financial support to all graduate students is extremely strong. 
This tradition is so pervasive that we were told by one administrator (a social 
scientist) that science students cannot pay their own way as do other graduate 
students. There is little difference in this commitment between the strong and the 
weak departments. One department with very little federal research money or 
university money for student support had reallocated a part of the equipment bud- 
get for student support. We found no strong trend toward students paying their own 
way. 

In the life science departments we visited, funds for training grants have been 
slightly declining in total dollar amount, even in strong departments. Federal fel- 
lowship support has almost disappeared from those departments. 

Student enrollments in the departments that have experienced declines in train- 
ing and fellowship funds have tended to decline, but not by as much as would be 
indicated by the extent of those drops alone. One reason is that most faculty we 

5 Unfortunately, data on undergraduate enrollments by field were not available for use in our quan- 
titative analyses of the determinants of faculty size. 



40 



interviewed on this subject strongly believed that significant reductions in graduate 
enrollment would not be good for their department, even if the only alternative is 
for faculty to put extraordinary effort into obtaining replacement funds. The factors 
explaining the strength of these beliefs were not entirely clear. They included 
faculty members' needs for assistance in research and a desire to have sufficient 
students around to make a broad teaching program worthwhile. Most department 
chairmen believed in the importance of teaching as a measure of faculty quality. 
Another factor is that undergraduate enrollment in some fields of biology is up 
substantially, creating a great need for graduate teaching assistants to avoid greatly 
increasing class sizes or faculty teaching loads. 

Few faculty thought that lack of job opportunities for graduate students was 
serious enough currently to be a reason for reducing enrollments in their depart- 
ments. On the contrary, most claimed that their students were finding jobs. In fact, 
chairmen of strong departments felt that their departments have a responsibility 
to continue training the faculty needed in other universities even if there is a 
nationwide surplus of graduate students. 

Faculty responses to cutbacks in training funds are highly varied. Some depart- 
ments, especially those receiving substantial federal research funds, shift students 
from training grant support to research assistantships. Several we saw had made 
concerted efforts to increase research grant support explicitly for the purpose of 
supporting students no longer covered by training grants and had been successful. 
Another response of some departments had been to cut back drastically on the 
number of postdoctoral students, which freed up funds from research grants to 
support graduate students. 

Some departments have approached their university administrations for addi- 
tional teaching assistantship positions. These are usually given to first and second 
year graduate students so that the reduced number of positions on training and 
research grants can be saved for advanced graduate students to finish their doctoral 
dissertations. The extent to which departments can obtain teaching assistantships 
varies greatly among universities and among departments within a university. As 
noted above, biochemistry is not taken by as many undergraduate students as 
general biology; consequently, biochemistry departments typically have far fewer 
teaching assistantships available for graduate student support than biology depart- 
ments. Thus, they are more highly dependent on federal support for graduate stu- 
dents. Also, teaching assistantships are more plentiful in state universities than in 
private universities. 

A small number of departments contemplated initiating or greatly increasing 
their masters degree program as a means of generating additional income. Our 
observation was that lower quality departments were likely to explore this alterna- 
tive. 

One university we visited has responded to the decline in federal student support 
by greatly increasing internal funds available for stipends and fellowships. Most 
universities cannot afford to take this approach. 

Three observations regarding graduate student enrollment emerge from this 
field work. First, the extent to which students in life science departments are pro- 
vided with financial support is dramatic. Second, departments are reluctant to 
reduce enrollments. They make a variety of adjustments to find support for students 
who can no longer be supported with federal training and fellowship funds. As a 



41 



result, enrollments do not decline in proportion to declines in federal support direct- 
ed to students. Third, departmental responses to cutbacks in funding are uneven. 
Departments have the greatest commitment to advanced graduate students; there- 
fore, cutbacks appear to fall most heavily on entering graduate students and post- 
doctoral students. 



FEDERAL FUNDING AND OUTSIDE FUNDING 

During our field trips, we explored relationships between federal research fund- 
ing and departmental ability to secure additional support from foundations, indus- 
try, and the state. In the universities we visited, neither industry nor foundations 
were a major source of direct support for research. 6 

Several of the department chairmen whom we interviewed had tried to ap- 
proach industry for support but without much success. It is hard for chairmen to 
identify which of the almost countless companies are likely to be receptive to pro- 
viding support, and who in those companies is likely to have sufficient authority to 
make decisions. By contrast, federal research agencies are organized specifically for 
the purpose of making grant awards to universities. Faculty generally know which 
of these agencies are likely sources of funding, how to apply, and what the decision- 
making process is. 

Ties with foundations are stronger, but since many will not pay the indirect costs 
of research, university administrators do not generally see them as an attractive 
source of funding. 



SCIENCE POLICY ISSUES 

During the field trips we were able to probe the opinions held by faculty and 
administrators at a diverse set of universities about several national science policy 
issues. Peer review received mixed reviews. Those who had done well in national 
competition tended to think it was an excellent system; those who had been rejected 
were conscious of its limitations. More to the point, the consensus among scientists 
is that although the system has its problems, it is a sound mechanism for assuring 
that quality ideas and research are rewarded with a minimum of political or other 
nonscientific interference. The major criticism we heard is that peer review in 
certain specialized areas requires critics who are bound to be competitors, in one way 
or another, of the proposed project director. 

The relative merits of project and institutional or center support were discussed. 
A number of scientists and administrators referred to the importance of Science 
Development and other institutional grants in the development of their more suc- 

6 Recently the American Council on Education, through its Higher Education Panel survey mech- 
anism, studied nonfederal funding of biomedical research and development at doctoral institutions. One 
survey question asked the institutional representatives to render a judgment about their expectations 
with respect to nonfederal funding of biomedical research at their institution during the next five years. 
"Only one-third of the respondents were anticipating significant increases. Public institutions tended to 
be slightly more optimistic than private institutions regarding an expansion in the nonfederal contribu- 
tions." F. J. Atelsek and I. L. Gomberg, Nonfederal Funding of Biomedical Research and Development: 
A Survey of Doctoral Institutions, Higher Education Panel Report #25, July 1975. 



42 



cessful departments. 7 Perhaps predictably, scientists tended to favor project support 
while administrators were quick to see the virtues of institutional support. 

In a number of institutions, great concern was expressed about the long run 
implications of the trend toward increased federal regulation of university policies 
that must be accepted with federal research funds. Regulation is rapidly spreading 
into new areas, including research on human subjects, student loan programs, free- 
dom of information, and requirements to publish data on the starting jobs and 
salaries of graduates. Universities are bearing the costs of compiling and submitting 
the data required to demonstrate compliance with these regulations, and the 
amounts are beginning to reach significant levels. Moreover, there is concern that 
as regulation increases, universities will become more subservient to the federal 
government and lose the institutional independence central to their function in 
society. 



UNCERTAINTY IN FEDERAL POLICY 

When asked to comment generally on the effects of federal funding policy, 
administrators, department chairmen, and faculty consistently responded that the 
ramifications of uncertainty and sudden shifts in federal policy (e.g., abrupt cessa- 
tion of programs and shifting of resources into new problem areas) were extremely 
difficult to cope with. One scientist described the problem as the reluctance of the 
federal government to provide enough continuity and assurance of support to last 
out the "intellectual lifetime of an idea." Another used the metaphor of the "unsta- 
ble patron" who distributes gifts that may be taken back unpredictably with serious 
consequences for the quality of work completed and the lives of individuals. Some 
scientists even went so far as to say that uncertainty in federal policy was more 
difficult to handle than predictably steady declines in funding levels. 8 

The problem is exacerbated by the peculiar complexity and structure of univer- 
sities as institutions. Three critical characteristics of universities are: (1) faculty 
have highly specialized skills, (2) a large proportion of them are tenured, and (3) 
departments make at least four-year commitments to graduate students. Because of 
these structural factors, adaptations to sudden changes in federal policy are painful 
and occur disproportionately in areas where there is greater flexibility. Two of these 
areas are the levels of untenured professors and entering graduate students. The 
effect is to upset the equilibrium among the many interrelated factors critical to the 
important functions of universities. Less sudden shifts in federal policy would allow 
universities to adapt more smoothly and maintain their equilibrium. 

7 For a description of the NSF Science Development program and an evaluation of its effects, see David 
E. Drew, Science Development: An Evaluation Study. National Academy of Sciences, Washington, D.C., 
1975. 

8 One pervasive influence of the uncertainty about federal policy is the reduced likelihood that a 
department will initiate some risky new venture with, for example, junior faculty. At several affluent 
and fairly undisturbed institutions, uncertainty in federal policy was cited, together with the leveling 
of funds, as causing the institution to reduce new ventures, experimental programs, etc. 



PART III 



Trends and Dimensions of Biomedical and Behavioral 
Research Funding in Academic Medical Centers 



Thomas E. Morgan and Daniel D. Jones 



Association of American Medical Colleges 



TABLE OF CONTENTS 

1. SUMMARY OF FINDINGS 1-1 

2. INTRODUCTION 2-1 

3. METHODS 3-1 

A. Data Sources 3-1 

B. Definition of Terms 3-4 

C. Adjustments for Inflation 3-6 

D. Categorization of Academic Medical 

Centers Research Involvement 3-8 

E. Index Numbers 3-13 

F. Time Serial Cross-Correlations 3-14 

4. RESULTS 4-1 

A. Trends in Academic Medical Center Funding 4-1 

1. NIH and NIMH Awards 1964-1974 4-1 

2. All Academic Medical Centers 4-5 

3. Established Academic Medical Centers 4-12 

4. Public Medical Centers 4-14 

5. Private Medical Centers 4-15 

6. Centers Ranked by Research Involvement 4-15 

7. Geographic Regional Differences 4-17 

8. New Medical Centers 4-18 

Tables and Graphs for Groups 4-21 

B. Special Studies 4-83 

1. Construction 4-83 

2. Organizational Complexity 4-85 

3. Curriculum Change and Federal 

Research Funding 4-87 

4. Changes in Distribution of NIH Awards 4-94 

5. Time Serial Correlations Between 

Funding Variables 4-94 

6. Patterns of Expenditures 4-97 

7. Administrative Arrangements for Biomedical 

Research 4-97 

8. Trends in Total Federal Revenues 4-98 

9. Career Choice of Medical Graduates 4-98 

5. DISCUSSION 5-1 

A. Summary of Findings 5-7 

B. Conclusions 5-12 

REFERENCES 



in 



LIST OF EXHIBITS 



Table 3-1 Price Deflators 3-7 

Table 3-2 Academic Medical Center Groupings 3-10 

Figure 4A-1A Health R§D as a Proportion of Total 4-2 

R§D Cost 

Figure 4A-1B NIH in Fiscal Year 1974 Funded... 4-2 

Figure 4A-1C NIH Obligated Funds by Functions 4-3 

Table 4A-1D NIH Grants to all Medical Centers 4-4 

Table 4A-1E Dollar Value of Selected NIH Programs 4-6 

Table 4A-1F Academic Medical Center Funding 4-7 

(Current $) 

Table 4A-1G Academic Medical Center Funding 4-8 

(Constant $) 

Figure 4A-1H Trends in Selected Academic Medical 4-9 

Center Finances 

Table 4A-1I Academic Medical Center Enrollment/ 4-11 

Faculty 

Table 4A-2A Funding in Established Academic 4-22 

Medical Centers 

Figure 4A-2B Average Medical Center Budget 4-23 

Table 4A-2C Biomedical and Behavioral Research 4-25 

Funding 

Table 4A-2D Academic Medical Center Enrollment 4-26 

and Faculty 

Figure 4A-2E Enrollment and Faculty 4-27 

Explanatory Notes for Group Tables 4-29 
and Graphs 

Table 4A-3A Public Medical Center Funding 4-30 

Figure 4A-3B Public Medical Center Funding 4-31 

Table 4A-3C Biomedical and Behavioral Research 4-33 

Funding in Public Medical Centers 

Table 4A-3D Public Medical Center Enrollment 4-34 

Figure 4A-3E Public Medical Center Enrollment 4-35 

Table 4A-4A Private Medical Center Funding 4-36 

Figure 4A-4B Private Medical Center Funding 4-37 

Table 4A-4C Biomedical and Behavioral Research 4-39 

Funding in Private Medical Centers 

Table 4A-4D Private Medical Center Enrollment 4-40 

Figure 4A-4E Private Medical Center Enrollment 4-41 

Table 4A-5A 1st Research Quartile Medical Center 4-42 

Funding 

Figure 4A-5B 1st Research Quartile Medical Center 4-43 

Funding 

v 



Table 4A-5C Biomedical and Behavioral Research 4-45 

Funding in 1st Research Quartile 

Medical Center 
Table 4A-5D 1st Research Quartile Medical Center 4-46 

Enrollment 
Figure 4A-5E 1st Research Quartile Medical Center 4-47 

Enrollment 

Table 4A-6A 2nd Research Quartile Medical Center 4-48 

Funding 
Figure 4A-6B 2nd Research Quartile Medical Center 4-49 

Funding 
Table 4A-6C Biomedical and Behavioral Research 4-51 

in 2nd Research Quartile Medical 

Centers 
Table 4A-6D 2nd Research Quartile Medical Center 4-52 

Enrollment 
Figure 4A-6E 2nd Research Quartile Medical Center 4-53 

Enrollment 

Table 4A-7A 3rd Research Quartile Medical Center 4-54 

Funding 
Figure 4A-7B 3rd Research Quartile Medical Center 4-55 

Funding 
Table 4A-7C 3rd Research Quartile Medical Center 4-56 

Enrollment 
Figure 4A-7D 3rd Research Quartile Medical Center 4-57 

Enrollment 

Table 4A-8A 4th Research Quartile Medical Center 4-58 

Funding 
Figure 4A-8B 4th Research Quartile Medical Center 4-59 

Funding 
Table 4A-8C 4th Research Quartile Medical Center 4-60 

Enrollment 
Figure 4A-8D 4th Research Quartile Medical Center 4-61 

Enrollment 

Table 4A-9A Northeastern Medical Center Funding 4-62 

Figure 4A-9B Northeastern Medical Center Funding 4-63 

Table 4A-9C Northeastern Medical Center Enrollment 4-64 

Figure 4A-9D Northeastern Medical Center Enrollment 4-65 

Table 4A-10A Southern Medical Center Funding 4-66 

Figure 4A-10B Southern Medical Center Funding 4-67 

Table 4A-10C Southern Medical Center Enrollment 4-68 

Figure 4A-10D Southern Medical Center Enrollment 4-69 

Table 4A-11A Midwestern Medical Center Funding 4-70 

Figure 4A-11B Midwestern Medical Center Funding 4-71 

Table 4A-11C Midwestern Medical Center Enrollment 4-72 

Figure 4A-11D Midwestern Medical Center Enrollment 4-73 

Table 4A-12A Western Medical Center Funding 4-74 

Figure 4A-12B Western Medical Center Funding 4-75 

Table 4A-12C Western Medical Center Enrollment 4-76 



VI 



Figure 4A-12D Western Medical Center Enrollment 4-77 

Table 4A-13A New Academic Medical Center Funding 4-78 

Figure 4A-13B New Academic Medical Center Funding 4-79 

Table 4A-13C New Academic Medical Center 4-80 

Enrollment 

Figure 4A-13D New Academic Medical Center 4-81 

Enrollment 

Table 4B-1 Federal Contribution to Academic 4-84 

Medical Center Construction 
Table 4B-2 Research Involvement and Organi- 4-86 

zational Complexity 
Table 4B-3A Federal Research Involvement and 4-89 

Curriculum Change 
Table 4B-3B Medical Scientist (M.D.-Ph.D.) 4-90 

Programs at Academic Medical 

Centers 
Table 4B-3C Family Medicine Programs 4-91 

Table 4B-3D Presence of Accelerated Programs 4-92 
Figure 4B-4 Change in Distribution of NIH Awards 4-95 
Figure 4B-5 Time Variation of Enrollment/ 4-96 

Employment with Federal Research 

Funding 
Table 4B-6 Percent of Total Academic Medical 4-99 

Center Operating Budgets from 

Federal Sources 
Table 4B-7 Comparative Number of Medical Graduates 4-101 

Entering Patient Care vs. Research 

and Teaching 

Table 5-1 Reconciliation of NIH Biomedical 5-6 

Support Data with AAMC Data 
Figure 5-2 Trends in Academic Medical Centers 5-15 



VII 



DIMENSIONS AND TRENDS IN HEALTH- RELATED RESEARCH FUNDING 

IN ACADEMIC MEDICAL CENTERS 



SUMMARY OF FINDINGS 



Between 1964 and 1974 the aggregate total operating budget of 
all academic medical centers increased 208 percent in current 
dollars and 82 percent in 1964 constant dollars. Federal re- 
search funding, on the other hand, increased 76 percent in 
current dollars and 6 percent in 1964 constant dollars. 

Indirect cost recoveries were closely related to total oper- 
ating budgets (up 2121 in current dollars). Due to inflation, 
indirect costs rose more rapidly than direct research costs. 
By 1974 indirect costs recovered from all sources equalled 
about 1/3 of tne direct costs of federally sponsored research. 

In 1965 federal funds for the direct costs of research accoun- 
ted for 36 percent of the total operating budget of 86 estab- 
lished academic medical centers. By 1974 this had fallen to 
23 percent. The percent decline was more marked for centers 
less involved in research and for those in the northeast and 
south. The percent decline in federal research funding was 
less marked for research-intensive and western centers. 

At the same time that federal biomedical research support as 
a fraction of total budget revenue was declining, academic 
centers received greater revenues from state and local govern- 
ments and from clinical and professional activities. State 
and local educational funds as a percentage of total operating 
budget increased from 13 percent in 1965 to 20 percent in 
1974. Professional fees and income from clinical activities 
increased from 4 percent to 23 percent over the same period. 
By 1974 each of these latter two categories accounted for ap- 
proximately the same dollar volume of revenue to the medical 
centers as did federal funds for direct costs of research. 

Federal funding instruments tended to shift over the ten-year 
period from investigator-initiated research grant to large, 
multi-investigator grants (program projects and centers) and 
contracts. Regular research grants as a fraction of total 
NIH awards declined slightly over the decade and the fraction 
of training grants, fellowships and faculty awards exhibited 
a more marked decrease. 

The data indicate that academic centers have increased greatly 
in complexity and in numbers of faculty and students of all 



1-1 



kinds. In the past ten years 30 new schools were established, 
and the remaining 86 schools showed a 471 increase in size and 
function. Overall, there was a doubling of graduate student 
enrollment and a marked increase in other health science stu- 
dent instruction. 

An examination of the possible impact of biomedical research 
funding on curriculum change indicated that, except for the 
larger number of medical scientist training programs (concur- 
rent award of M.D./Ph.D. degrees) in research-intensive and 
private institutions, there was no relation between biomedical 
research funding and curriculum development. However, many 
schools had undergone extensive curriculum changes, shortening 
medical education programs and initiating new programs for the 
training of family medicine and other specialized types of phy- 
sicians. Research "intensive" schools started as many family 
medicine programs as schools less involved in research. 

Although large expenditures have been made for biomedical re- 
search in medical centers over the past twenty years, we were 
able to find no relationship between the complexity of academic 
centers and increased biomedical research funding except in the 
first quartile of research-intensive academic centers. These 
institutions had established more health science research and 
development centers in response to the availability of biomedi- 
cal research funds. They also had as many hospital affilia- 
tions as did institutions less involved in research. 



1-2 



2. INTRODUCTION 

In June, 1975, the Association of American Medical Colleges 
contracted with the President's Biomedical Research Panel as part 
of a consortium with the American Council on Education and the 
Rand Corporation to study aspects of biomedical research in the 
United States. The task of the Association was to depict the di- 
mensions and trends of biomedical and behavioral research funding 
to the academic medical centers of the United States. The spe- 
cific questions about which the Association was to gather data 
from 1964-74 and report are as follows: 

1. What is the total operating budget of each academic 
medical center and what part is provided by student 
fees, capitation and special project grants, state 
and local government funds, endowments and gifts, re- 
search funds, service programs, indirect cost reimburse- 
ments, professional fees, general research support grants 
and other income? 

2. What trends in funding have occurred for the trans- 
fer of NIH and ADAMHA funds to academic medical 
centers through regular research grants, program 
project grants, center grants, contracts, train- 
ing grants, faculty awards, general research grants, 
clinical research centers, construction and reno- 
vation grants and loans and National Library of 
Medicine awards? 



2-1 



3. Between the years 1964 through 1974, what changes 
have there been in faculties and students in total 
student enrollments, medical and graduate student en- 
rollment, faculty size, type and funding? 

4. What part of funding of new construction and renovation 
has been provided by federal, non-federal sources? 

5. In what ways has NIH/ADAMHA funding of centers, other 
research and research training affected the medical 
school curriculum in recipient schools? 

6. How do organizational structures for the adminis- 
tration of NIH, ADAMHA, and other federal research 
funding vary? 

7. What functions are performed by special administrative 
units in the schools created for the purpose of manag- 
ing and expanding federal biomedical research funding? 

During the six months available for the Study the data was acquir- 
ed, loaded onto computer files and verified. Because the raw data 
consisted of more than 900 pages of computer output in tabular 
form, our approach to simplification of the data has been to de- 
pict graphically, trends in the finances of academic centers over 
the decade 1964-1974. Major revenue sources have been shown and 
all revenues have been adjusted for inflation. 

An additional charge was to estimate the "impact" of research 
funding on medical centers, therefore, for reasons of comparabil- 
ity the major portion of the study concerns only those schools 

1/ 
established prior to 1958. In addition, however, sections on 

new and emerging schools have been included. 



1_/ No new schools were established between 1956 and 1960 

2-2 



In order to achieve a further reduction of the data, academic 

medical centers were grouped by 1) public/private ownership, 
2) research emphasis and 3) geographic distribution. Research 
emphasis was determined by a factor technique described in the 
Methods Section. Average revenues and/or federal awards for each 
group are presented. A number of other approaches were considered 
and rejected after consultation with staff and an ad hoc advisory 
group . 

As the trend and dimensional data were collected, we attempted 
to consider ways in which federal research funds may have affected 
academic medical centers. In this regard, we attempted to estimate 
whether large amounts of federal research funds are associated with 
organizational complexity of the centers, with faculty size, com- 
position or salaries, with construction of facilities, with change 
in teaching or service program emphasis, with curriculum innova- 
tion or with increased dependency on research funds. In general, 
Task 2 of the overall Study confines itself to a reporting of the 
facts. However, whenever possible, trends and conclusions found 
by other contractors studying the Universities and departments in 
selected medical schools were compared with our own findings. 
These more general conclusions are summarized in Task 5. 



2-3 



3. METHODS 

A. DATA SOURCES 

Two principal sources of data were used for the Task 2 study. 
The first was the Institutional Profile System (IPS) , a computer- 
ized data base created and maintained by the Association of Ameri- 
can Medical Colleges. The second major data source was the Infor- 
mation for Management Planning Analysis and Coordination (IMPAC) 
file created and maintained by the Division of Research Grants at 
the National Institutes of Health. Both data sources are describ- 
ed briefly below. 

1 . Institutional Profile System (IPS) 

This system was established by the AAMC in 1973 in order to 
improve the accuracy, speed and comprehensiveness of the collec- 
tion, storage and retrieval of data relating to academic medical 
centers. IPS serves the dual purpose of meeting many of the in- 
formation needs of the academic medical centers and also of func- 
tioning as a powerful information tool for external users. The 
data base now contains in excess of 6,000 data items describing 
many aspects of the finance, enrollment, faculty, curriculum and 
service programs of each medical school. Since its inception in 
1973, the IPS has been updated with newly collected data and ex- 
tended backward in time as far back as 1960 for many previously- 
collected data items. 

The principal data collection instrument for the IPS is the 
annual questionnaire of the Liaison Committee on Medical Education 

3-1 



(LCME) (1), a joint effort of the AAMC and the AMA. The LCME ques- 
tionnaire consists of two parts, Part I of which covers the reve- 
nues and expenditures for current operations of a medical school 
during a given fiscal year. Part II covers student enrollment, 
faculty employment, curriculum, construction and student financial 
aid. Because the scope and content of the LCME questionnaires 
have changed over the years, special measures were required for 
Task 2 to facilitate compatibility and internal consistency of the 
data in the time domain. Accordingly, detailed maps were construc- 
ted relating comparable items between different editions of the 
questionnaire. The correspondences were not always one-to-one, 
there being an appreciable number of one-to-many and many-to-many 
correspondences as well as some items which had no correspondence 
with any items on other editions of the questionnaire. In terms 
of the items of information required for successful completion of 
the present Task, we and our financial consultants feel that the 
mapping scheme actually employed by us represents the optimal 
solution to the correspondence problem. 

Hardware, software and file design : The IPS data base is 
housed on disc-packs operating on a Xerox Sigma 6 computer. Re- 
trieval capability in a multi -processing , time-sharing environment 
is provided by an AAMC-written data management software package 
with remote, on-line terminal access. The data is organized by in- 
stitution and by variable number with variable descriptions also 
stored on disc. Corresponding variables for successive years may 
be chained together by suitable pointers to obtain a discrete time 
series. The IPS software package provides numerous capabilities 
such as listing, ranking, printer-plotting, correlation and real- 
time computation of new variables. 

3-2 



2. NIH/IMPAC File 

The primary objective of the NIH/IMPAC File is the establish- 
ment and maintenance of a data base from which complete, accurate 
and timely reports to all levels of federal management can be gen- 
erated. The IMPAC file is maintained by the Statistics and Analy- 
sis Branch, Division of Research Grants, National Institutes of 
Health. It contains principally data on NIH Extramural Programs 
supplemented by data on research, training and manpower education 
programs of other bureaus of the Public Health Service. The con- 
tents of the IMPAC File are detailed in a series of IMPAC Infor- 
mation and Instruction Handbooks, e.g., (2). 

The aggregation of NIH program types used in the present study 
is slightly different from the standard designations used by NIH. 
NIH programs were aggregated as follows: 

Program Type IMPAC Activity Codes 

Regular Research Grants R01 through R28 

Program Project Grants P01, P06, P07, P09, P13, P15 

Center Grants P10, Pll, P16, P17, P18 

Clinical Research Centers M01 , P02 

General Research Grants SOI through S06 

Training Grants T01 through T41 

Fellowships F01 through F22 except F14 

Contracts N01 

The AAMC requested and received a subset of the IMPAC file 

from the Division of Research Grants, NIH. The data subset was 

arrayed by agency, program, year and recipient institution using 

the standard major component search code combination of 01, 05 and 

12. Aggregated summary tables for NIH, NIMH and BHME as well as 

an overall summary table were provided. The media on which the 

IMPAC subset was written were hardcopy and magnetic tape. 



3-3 



B. DEFINITION OF TERMS 

The definitions which have been used in addressing the ques- 
tions posed in the biomedical and behavioral research study (Task 
2) are as follows: 

"Total operating budget" is defined as total current funds ex- 
penditures as reported in the LCME Questionnaire Part I, line 66. 
This definition rather than total current funds revenues was chosen 
because it is a better indication of the total financial burden on 
the academic medical centers. A consequence of this choice is that 
the sum of the revenue components will occasionally exceed the to- 
tal operating budget, particularly for new and developing medical 
schools . 

"Student fees" is equated to student tuition and fees with 
the recognition that this amount is a revenue figure which in some 
institutions is estimated rather than precisely known. 

The amounts of federal capitation grants could not be deter- 
mined accurately and consistently from the LCME Questionnaire. The 
primary reason for this is that the reporting schools variably con- 
sider capitation as unrestricted revenue. Thus, they have the op- 
tion of commingling capitation grant funds with other income avail- 
able for regular operating costs of various programs. Different 
schools elected different options for reporting capitation grant 
funds. Moreover, federal capitation grant funds were not reported 
separately prior to 1971 before which they were designated Basic 
Improvement Grants. For these reasons, data on federal capitation, 
basic improvement and special project grants were obtained from a 
Bureau of Health Manpower document, e.g., (3). It must be remem- 
bered that these figures for federal capitation, basic improvement 

and special project grants represent funds awarded rather than 
funds expended during a given time interval. 

3-4 



"State and local government funds" are defined as the sum of 
state school appropriations, special appropriations, private school 
subsidies, interstate and/or intrastate compacts, city and county 
government funds and state and local government-sponsored research. 

"Endowments and gifts" are defined as only unrestricted en- 
dowments and gifts. 

"Non-government research funds" are defined as the sum of the 
non-government sponsored research funds and other separately bud- 
geted research funds. 

"Service programs" are defined as sponsored multi-purpose and 
service programs. This category includes restricted revenues from 
governmental agencies and private sources which support health ser- 
vices or institutional grants covering health services and other 
of the institution's basic functions. Among the reportable reve- 
nues in this category might be health service contracts with hos- 
pitals or clinics, Regional Medical Programs and health science 
library resources. 

"Indirect cost" reimbursement is defined as indirect cost re- 
covery resulting from sponsored research, sponsored teaching and 
training and sponsored multi-purpose and service programs. 

"Professional fees" are defined as revenue generated through 
medical service programs in which income accrues to the medical 
school for the support of its programs. 

The amount of general research grants could not be determined 
from the LCME Questionnaire and was therefore obtained from the 
NIH/IMPAC file instead. 

"Miscellaneous income" is defined as the sum of revenues re- 
sulting from the sales and services of educational departments, 

organized activities related to educational departments, and other 

sources . 

3-5 



The "other" category appearing in certain graphs in the pre- 
sent report does not correspond to the "miscellaneous" category 
defined above. The magnitude of the "other" category was deter- 
mined by subtracting the sum of federal sponsored research, federal 
teaching and training, federal capitation and special projects, 
state and local government funds, endowments and gifts, service 
programs, professional fees, and non-federal sponsored research 
from the total operating budget. 

"Total student enrollment" is defined as the sum of medical 
student enrollment, graduate student enrollment, house staff (in- 
terns and residents), post-doctorals and medical student equiva- 
lents as defined by the AAMC. (1) 

C. ADJUSTMENTS FOR INFLATION 

In order to facilitate comparisons of financial variables in 
the time domain, the data used in the present study were adjusted 
for inflation. Due to the variety of purposes for which various 
monies were expended and received, several different price indices 
were employed to effect this adjustment. The four different price 
indices are displayed in Table 3-1 and described briefly below. 
1 . NIH Biomedical Research and Development Price Index 
A series of biomedical R § D deflators have been developed 
under a contract awarded by NIH. (5) Of these, the biomedical R 
§ D price index for the academic sector was used to deflate both 
federal research funds as reported by the medical schools and re- 
search grants and contracts awarded to medical schools by NIH. 
The NIH R § D price index covers direct costs only. It is a weighted 
average of two price indices, one for personnel compensation and 
one for non-personnel costs. The major budget categories are: 



3-6 



TABLE 3-1 



PRICE DEFLATORS USED IN CONVERTING CURRENT 
DOLLARS TO CONSTANT DOLLARS 





Halstead 


Halstead 








Higher Educ. 


Halstead 1 


NIH 2 




Instruction 


Construction 


Equipment 


R$D 


Fiscal 


Price 


Price 


Price 


Price 


Year 


Index 


Index 


Index 


Index 



1964 100.0 100.0 100.0 100.0 

1965 104.3 103.0 

1966 109.5 106.8 

1967 115.2 111.9 

1968 122.1 120.0 

1969 130.4 129.2 

1970 139.4 138.7 

1971 148.3 150.7 

1972 156.3 163.0 

1973 164.5 173.2 

1974 176.0 184.9 



100.2 


102.6 


101.3 


106.5 


105.2 


111.2 


108.4 


117.6 


111.3 


124.0 


116.1 


130.9 


121. 5 


138.5 


124.0 


144.3 


127.8 


150.5 


137.8 


160.2 



See reference 6, 
See reference 5, 



personal costs, including consultants, supplies, equipment, patient 
care, travel, alterations and others. The major categories under 
personnel compensation are: faculty members, research associates, 
postdoctorals , graduate students, other professionals, research 
technicians, clerical and secretarial, and other support staff. 
The construction of the NIH R § D price index is described in fur- 
ther detail by Jaffe and Adelman (5). 



3-7 



2 . Halstead Instruction and Construction Price Indices 
D. Kent Halstead has developed a number of price indices for 
various activities related to higher education (6). Two of these, 
the instruction price index and the construction price index, were 
used to deflate certain financial data in the present study. Per- 
sonnel compensation accounts for 82 percent of the instruction 
price index and includes professional salaries, non-professional 
wages and salaries, and fringe benefits. The remaining 18 percent 
of the instruction price index covers services, supplies and ma- 
terials, equipment, books and periodicals and utilities. The in- 
struction price index was used to deflate medical school financial 
data excluding sponsored research and construction. 

The construction price index described by Halstead is essen- 
tially that of the Boeckh Division of the American Appraisal Com- 
pany. It is a weighted average of building construction and movable 
equipment. The building construction index covers wage rates and 
structural building materials including plumbing, heating, lighting 
and elevators. The movable equipment index covers commercial fur- 
niture and machinery and equipment for office and general uses. 
In the present study, the construction price index was used to de- 
flate the cost of construction completed in a given year. Construc- 
tion initiated or planned was not included in the present study. 
Both the instruction price index and the construction price index 
are described in detail by D. Kent Halstead (6). 

D. CATEGORIZATION OF ACADEMIC MEDICAL CENTERS RESEARCH 
INVOLVEMENT 

In order to effect a concise and parsimonious presentation of 
financial/enrollment data on 100-odd medical schools over a 10- 
year period, it was decided to group the schools into aggregates 
that might reflect the unique impact of changes in federal research 

3-8 



funding. Toward this end, three variables were defined and com- 
bined into a composite research orientation factor using the FAC- 
TOR program of the Statistical Package for the Social Sciences (4) , 
The three input variables were: 

(1) federal research dollars/total operating budget 

(2) federal research dollars/full-time faculty 

(3) federal research dollars/basic science faculty 

The specific method selected was principal factoring without iter- 
ation leaving the main diagonal of the correlation matrix unalter- 
ed. This method extracts principal components without requiring 
any assumptions about the general structure of the variables. As 
it turned out, the weights obtained from the principal component 
extraction were not appreciably different from the assignment of 
equal weights to the three variables. The net effect of this out- 
come is to weight the faculty research intensiveness about twice 
that of the dollar research intensiveness. The value of the re- 
search orientation factor for each academic medical center was 
then calculated using the standardized values of the three varia- 
bles defined above and the corresponding weights derived from the 
principal component extraction. The academic medical centers 
were then ranked in descending order of their research orientation 
and divided into quartiles from which the research quartiles were 
derived. 

It should be emphasized that the ranking is heavily dependent 
on the research and total budget revenues for a single year, 1972. 
Also, differences between adjacent quartiles were quite small and 
the research factor, in adjusting for size of the institution, was 
strongly affected by non-research revenues. For example, an insti- 
tution with a large budget derived equally from research, teaching 
and clinical activities would be likely to rank lower than a small 

3-9 



TABLE 3-2 



ACADEMIC MEDICAL CENTER GROUPINGS 
86 Established Medical Centers 



Private (43) 



Type of Ownership 

Public (43) 



Albany 

Baylor 

Boston 

Bowman- Gray 

Chicago Medical 

U. Chicago 

Columbia 

Cornell 

Creighton 

Dartmouth 

Duke 

Einstein 

Emory 

Georgetown 

George Washington 

Hahnemann 

Harvard 

Howard 

Jefferson 

Johns Hopkins 

Louisville 

Loyola 

Miami 

Medical College of Wisconsin 

Loma Linda 

Meharry 

New York Medical 

New York Unviersity 

Northwestern 

U. Pennsylvania 

Pittsburgh 

Rochester 

St. Louis 

Southern California 

Stanford 

Temple 

Tufts 

Tulane 

Vanderbilt 

Washington U. St. Louis 

Case Western 

Medical College of Pennsylvania 

Yale 



Alabama 

Arkansas 

SUNY Buffalo 

California, San Francisco 

California, Los Angeles 

Cincinnati 

Colorado 

Florida 

Georgia 

Illinois 

Indiana 

California, Irvine 

Iowa 

Kansas 

Kentucky 

SUNY Downstate 

Louisiana, New Orleans 

Maryland 

U. Michigan 

Minnesota, Minneapolis 

Mississippi 

Missouri, Columbia 

Nebraska 

North Carolina 

North Dakota 

Ohio State 

Oklahoma 

Oregon 

South Carolina 

South Dakota 

Texas , Southwestern 

New Jersey 

SUNY Upstate 

Tennessee 

Texas, Galveston 

Utah 

Vermont 

U. Virginia 

Medical College of Virginia 

U. Washington, Seattle 

Wayne State 

West Virginia 

Wisconsin 



3-10 



Northeast (28) 



TABLE 3-2 (Continued) 
Geographical 

South (25) 



Albany 

Boston 

SUNY Buffalo 

Columbia 

Cornell 

Dartmouth 

Einstein 

Georgetown 

George Washington 

Hahnemann 

Harvard 

Howard 

Jefferson 

Johns Hopkins 

SUNY Downstate 

Maryland 

New York Medical 

New York University 

U. Pennsylvania 

Pittsburgh 

Rochester 

New Jersey 

SUNY Upstate 

Temple 

Tufts 

Vermont 

Medical College of Pennsylvania 

Yale 



West (10 ) 

California, San Francisco 

California, Los Angeles 

Colorado 

California, Irvine 

Loma Linda 

Oregon 

Southern California 

Stanford 

Utah 

U. Washington, Seattle 



Alabama 

Arkansas 

Baylor 

Bowman-Gray 

Florida 

Duke 

Emory 

Georgia 

Kentucky 

Louisiana, New Orleans 

Louisville 

Miami 

Meharry 

Mississippi 

North Carolina 

Texas Southwestern 

Tennessee 

Texas Galveston 

Tulane 

Vanderbilt 

U. Virginia 

Medical College of Virginia 

West Virginia 



Midwest (2 3 ) 

Chicago Medical 

U. Chicago 

Cincinnati 

Creighton 

Illinois 

Indiana 

Iowa 

Kansas 

Loyola 

Medical College of Wisconsin 

U. Michigan 

Minnesota 

Missouri 

Nebraska 

North Dakota 

Northwestern 

Ohio State 

St. Louis 

South Dakota 

Washington U. , St. Louis 

Wayne State 

Case Western 

Wisconsin 



3-11 



TABLE 3-2 (Continued) 



New Medical Schools Synchronized by Year of Operation (10 ) 

Arizona 

Brown 

California/Davis 

California/San Diego 

Connecticut 

Mt. Sinai 

Michigan State 

Penn State 

Rutgers 

Texas/San Antonio 



3-12 



institution with a larger proportion of research revenues. 

Other categorizations : In addition to the research involve- 
ment categories, other more traditional groupings were employed. 
These include: 

(1) type of ownership (public, private) 

(2) geographic location (northeast, south, midwest, west) 

(3) year of founding (schools founded before/after 1958) 

These groups of schools are also enumerated in Table 3-2. 

E. INDEX NUMBERS 

Academic medical centers in the United States cover a wide 
range of magnitude both in terms of students and in terms of total 
operating budget. For example, total student enrollment in 1974 
ranged from 24 to 3,042 and medical center operating budgets ranged 
from $1.3 million to $79 million. Descriptions of comparative di- 
mensions and trends within such a heterogeneous population can be 
very misleading if the size factor is not taken into account. In 
order to facilitate the comparative description of dimensions and 
trends in medical center funding and enrollment, the method of in - 
dex numbers was selected for use in the graphical displays. Index 
numbers are widely used by educators in describing trends in fund- 
ing and enrollment. In the application of this method, the value 
of a given financial or enrollment variable in an arbitrarily chos- 
en reference year is equated to 1.0. The values of the variables 
for the other years relative to the reference year are then com- 
puted and displayed accordingly. The calculation of index numbers 
is completely independent of other adjustments such as those made 
for inflation, i.e., the base years for index numbers and constant 
dollars do not have to be the same. Index numbers have the advan- 
tage that variables of widely differing magnitude can be displayed 
on the same graph and between-graph comparisons can be made in a 

3-13 



straightforward manner. In the present report, the absolute value 
corresponding to the reference year is usually given so that quick 
conversions from index number to absolute values can be made read- 
ily. In any case, the absolute values of all variables are avail- 
able in the appropriate tables. 

F. TIME SERIAL CROSS -CORRELATIONS 

It is likely that certain effects of federal research funding 
on academic medical centers are realized only after a delay of a 
year or more. In order to investigate such delayed effects on fin- 
ancial and enrollment variables, the data were analyzed by the 
methods of time-series analysis. The specific technique used was 
the calculation of cross-correlation coefficients between selected 
pairs of variables for time lags of 0, 1, 2 and 3 years. The basic 
input data consisted of the annual means of various financial and 
enrollment variables averaged over 86 established medical schools 
for durations ranging from 10 years to 14 years depending on the 
availability of data. In order to circumvent spurious correlations 
due to concurrent long-term trends in the data, the raw time series 
data were passed through a first-order difference filter. After 
detrending the data in this manner, the cross-correlation coeffi- 
cient between variable x and variable y was calculated from the 
equation: 



c(u) — 



n-u 



I 

JliL 



(x.j " *) (yj +u - y) 




x) 2 



(y-i - 7) 



3-14 



where 

c is the cross-correlation coefficient 

u is the time lag in years 

n is the record length in years 

x is the mean of x 

y is the mean of y 
and j is the index of summation within the indicated limits 

As an aid in assessing the significance of the cross -correlation 
coefficients so obtained, the sample "t" statistic was computed 
from 



V^ 



V 



c 2 



where c is the sample cross -correlation coefficient and n is the 
sample size. Under the assumptions of normal correlation analysis, 
the sample t statistic is distributed as student's t statistic 
with n-2 degrees of freedom and it can be used to test the null 
hypothesis that the population cross-correlation coefficient equals 
zero . 



3-15 



SECTION 4: RESULTS 

A. TRENDS IN ACADEMIC MEDICAL CENTER FUNDING 

1. NIH and NIMH Awards 1964-1974 

Between 1964 and 1974, health research and development spend- 
ing in the United States grew both in absolute dollars and as a 
proportion of total research and development costs. This is illus- 
trated in Figure 4A-1A where it can be seen that health research 
and development (R § D) increased from $1.6 billion to $4.3 billion 
and from 8.6 percent to 13.3 percent of total R § D costs. Figure 
4A-1A also shows that in 1974 65 percent of all health R § D sup- 
port came from the federal government and that approximately 36 
percent of all health R § D was performed in an academic environ- 
ment . 

In 1974, the National Institutes of Health (NIH) were the lar- 
gest single source of federal support for academic R 5 D, This is 
shown in Figure 4A-1B where it can be seen that although NIH ac- 
counted for only 10 percent of all federally supported R § D, it 
supported 46 percent of all federally sponsored R § D at academic 
medical centers. Figure 4A-1C depicts the twenty year growth of 
NIH obligated funds in current dollars. (Figures 4A-1A through 
1C were taken from "Basic Data Relating to the National Institutes 
of Health: 1975.) (7) 

If economic inflation is taken into account, a different time 
profile of NIH R § D support emerges. Table 4A-1D shows NIH awards 
to academic medical centers by program from 1968 to 1974. 

4-1 



Figure 4A-1A 

HEALTH R it D AS A PROPORTION OF TOTAL R&D COST 

UNITED STATES 1964 AND 1974 EST 



TOTAL R&D COST 




SUPPORT OF HEALTH RbO. 1974 EST 
BY SOURCE BY PERFORMER 





54 3 BILLION 

HEALTH RESEARCH AND DEVELOPMENT 

51 6 BILLION - 8 6% OF 1964 TOTAL REfD 



54 3 BILLION 




54 3 BILLION* 13 3% OF 1974 TOTAL R&D 



SOURCE FOR TOTAL R&D NATIONAL SCIIHCC fOUNDATION FOR HEALTH R«,D NIH 

* $2.7 BILLION 1964 CONSTANT DOLLARS 



Source : "Basic Data Relating to the National Institutes of 
Health: 1975", W.T. Carrigan, Ed. 



Figure 4A-1B 

NIH IN FISCAL YEAR 1974 FUNDED. 



10% OF ALL FEDERALLY 
SUPPORTED R&D 




ER DHEW 



517 7 BILLION 



SOURCES NSF F EDERAL FUNDS FOR RESEARCH DEVE LOPMENT. 
ANO OTHER SCIENTIFIC ACTIVITIES. VOL XXIII, TABLE re 
AND NIH FOR DATA ON NIH, REFLECTED IN TOTALS 



FEDERAL SUPPORT FY11MEST 
(IN MILLIONS!* 



AGENCY 


aii Ff n 

FiAO 


FlftD AT 

ACAD 
INST 


TOTAL 


SW.7-T3 


S? ?3f, 


IIMFW 


? 33^, 


1 ?30 


NIH 


1 738 


1 (114 


OTHER 


&HR 


lOfi 


NSF 


S30 


418 


OOD 


8 f -99 


19? 


NASA 


3 076 


88 


AEC 


1 -131 


84 


OTHER 


1,81 1 


224 



46% OF FEDERALLY SUPPORTED R&D 
AT ACADEMIC INSTITUTIONS 




$2 2 BILLION 



•INCLUDES FUNDS WITH HELD IN FY 1973 
AND RELEASED FOR USE IN 19;a 



Source: "Basic Data Relating to the National Institutes of 
Health; 1975", W.T. Carrigan, Ed. 



4-2 



FIGURE 4A-1C 



MILLIONS 
$2000 i— 



1000 



500 



T — i 1 r 



NIH OBLIGATED FUNDS BY FUNCTION, FY 1955- 1975 ESTIMATE 

EXCLUDING PROGRAMS THAT HAVE BEEN TRANSFERRED OUT' 

™ EST. TOTAL NIH 

« FY 1975 $1,757 




PROGRAM M 



RESEARCH 

CONSTRUCTION 

GRANTS 



1 955 



FY 1975' 



• N'MII COMMUNITY PROGRAMS PBS AND RMUE A I SO E X CHIDE S FOREIGN CUR RE NCY PROGRAM AND NIMH rolU ION OF GRS GRAN I S 
INCLUDES NLM FROM FY I9G8 ON tlNCLUDESRUILDINGS AND FACILITIES FROM FY 1969 ON BUT CHART EXCLUDES DIRECT CONSTRUC 
TION FOR PRIOR YEARS | INCl UOES $77? MILLION IN FY 1973/74 RELEASED FUNDS * FY 1 9 75 COLUMN OF 1976PRFSIDENTS BUDGET 
IREOSION LEVEL! 



NIH obligations include indirect costs and research grants to per- 
formers other than academic medical centers. NIH obligations precede 
medical center reports of expenditures by at least a year. 

SOURCE: "Basic Data Relating to the National Institutes of Health: 
1975", W.T. Carrigan, Ed. 



4-3 



TABLE 4A-1D 
NIH GRANTS TO ALL MEDICAL CENTERS 1 



Totals in Thousands of 1964 Constant Dollars 



Fiscal Year 



Program 


1968 


1969 


1970 


1971 


1972 


1973 


1974 


Regular Research 
Grants 


165, 446 3 


155,404 


139,094 


142,834 


161,088 


160,311 


198,254 


Program Project 
Grants'* 


49,917 


"53,075 


49,906 


57,791 


71,840 


75,562 


90,532 


Center Grants^ 


3,776 


4,198 


4,585 


14,752 


20,136 


22,241 


23,951 


Contracts 


30 


12 


22 


47 


2,737 


50,952 


70,765 


Training Grants 


72,776 


72,776 


64,322 


61,585 


62,913 


46,506 


63,165 


Faculty Awards 


16,454 


16,050 


14,757 


13,778 


12,772 


11,268 


9,968 


General Research 
Grants 


23,366 


22,551 


20,290 


18,758 


17,164 


6,705 


22,634 


Clinical Research' 


29,849 


30,879 


29,569 


29,208 


32,565 


31,779 


32,000 


Construction Grants 


31,372 


6,701 


18 





9,044 


12,166 


9,067 


NLM Awards 


1,832 


1,921 


1,415 


786 


919 


944 


1,052 


Fellowships 


9,134 


9,590 


7,417 


7,746 


6,881 


3,734 


9,612 


NIH TOTAL 


403,953 


373,147 


331,141 


347,284 


388,059 


402,059 


531,000 



'Source: NIH IMP AC File. 
Constant dollars computed with NIH R & D. deflator. 

Program totals may not sum to NIH total because not all programs are shown. 

4 
Program project grants and center grants are aggregated differently here than in the standard 

NIH program classification. Program project grants include the IMPAC activity codes POl , P06, 

P07, P09, P13 and P15. Center grants include P10, Pll, P16, P17 and P18. Clinical research 

centers include MOl and P02. 



4-4 



The basic data were obtained from the NIH/IMPAC file and deflated 
using the NIH R § D Price Index (see Methods Section for details). 
Between 1968 and 1970 there was a 73 million dollar decrease in 
constant dollar awards to academic medical centers. This corres- 
ponds to an 18 percent drop. There followed a slow rise in funds, 
but only in 1974 did constant dollar funding exceed the 1968 
level. Funding in 1974 was 311 above the 1968 level. 

Program mechanisms by which NIH funds are transferred to the 
academic medical centers in support of the biomedical and behavior- 
al research efforts were examined in detail. Regular research 
grants, the traditional principal mechanism for the transfer of 
funds, followed the same general trend as total awards. However, 
after the general decline in funds in fiscal 1969, regular research 
grants awards did not recover as well as total awards so that reg- 
ular research grant funding in fiscal 1974 was only 20% above the 
1968 level. This was due primarily to a 132 million dollar gain 
(245%) in other types of awards -- center grants, program project 
grants, and contracts. As shown in Table 4A-1E, all other funding 
programs were either stable or declined slightly during the decade. 
2. All Academic Medical Centers (N=*86 to 116) 

The intention of this study is to discuss trends in biomedical 
and behavioral research funding from the perspective of the academ- 
ic medical centers. This should serve to complement the more com- 
monly stated viewpoint of the Federal provider of funds. This ob- 
jective is highlighted in exhibits 4A-1F through 4A-1H which show 
the total revenues of the academic medical centers over the past 
10 years. These revenues are presented both as current dollars 
and as 1964 constant dollars for all medical centers for which data 
is available. Overall, between 1964 and 1974, academic medical 

4-5 



TABLE 4A-1E 

DOLLAR VALUE OF SELECTED NIH PROGRAM AWARDS AS A PERCENT 
OF TOTAL NIH AWARDS 1 



Program 



Fiscal Year 
1968 1969 1970 1971 1972 1973 1974 



Regular Research „ 

Grants 41% 2 42% 42% 41% 40% 40% 37% 

Program Project 

Grants 12% 14% 15% 17% 18% 19% 17% 

Center Grants 3 1% 1% 1% 4% 5% 6% 4% 

Contracts 0% 0% 0%' 0% 1% 13% 13% 

Training Grants 18% 20% 19% 18% 16% 12% 12% 

Faculty Awards 4% 4% 4% 4% 3% 3% 2% 

Fellowships 2% 3% 2% 2% 2% 1% 2% 



1 Source: NIH IMPAC file. 



9 

Percentages will not sum to 10 because not all programs are shown. 

3 See footnote 4 Table 4A-1D. 



4-6 



FIGURE 4A-1F 



ACADEMIC MEDICAL CENTER FUNDING, 1964-1974 

(Total revenues for all centers in 
millions of current dollars) 



Fiscal Year 



1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 



Total Operating 
Budget 

Federal Research 
Funds 

State & Local 
Government Funds 

Professional Fees 
& Service Programs 

2 
Indirect Costs 



Federal Teaching 
& Training- 3 

Federal Capitation 



Non-Federal 
Research 

Student Fees 



Endowments and 
Gifts 

Other 5 



783 


881 


1015 


1173 


1181 


1552 


1770 


2020 


2171 


2409 


280 


307 


347 


388 


338 


386 


376 


441 


432 


493 


36% 


35% 


34% 


33% 


29% 


25% 


21% 


22% 


22% 


21% 


100 


116 


143 


156 


191 


235 


296 


260 


393 


471 


13% 


13% 


14% 


13% 


16% 


15% 


17% 


13% 


18% 


20% 


31 


41 


51 


87 


139 


283 


346 


388 


466 


522 


4% 


5% 


5% 


7% 


12% 


18% 


20% 


19% 


21% 


22% 


49 


58 


60 


82 


82 


96 


105 


125 


149 


153 


6% 


7% 


6% 


7% 


7% 


6% 


6% 


6% 


7% 


6% 


96 


111 


127 


153 


136 


168 


196 


219 


245 


251 


12% 


13% 


13% 


13% 


12% 


11% 


11% 


11% 


11% 


10% 


.A. 


N.A. 


7 


19 


26 


41 


56 


77 


131 


125 






1% 


2% 


2% 


3% 


3% 


4% 


6% 


5% 


43 


48 


52 


56 


52 


73 


87 


31 


117 


44 


5% 


5% 


5% 


5% 


4% 


5% 


5% 


2% 


5% 


2% 


38 


41 


43 


47 


44 


55 


63 


78 


92 


106 


5% 


5% 


4% 


4% 


4% 


4% 


4% 


4% 


4% 


4% 


33 


37 


38 


37 


42 


47 


52 


67 


71 


63 


4% 


4% 


4% 


3% 


4% 


3% 


3% 


3% 


3% 


3% 


113 


122 


147 


148 


131 


168 


193 


334 


75 


181 


14% 


14% 


14% 


13% 


11% 


11% 


11% 


17% 


3% 


8% 



Sources: AAMC Financial Questionnaire and Bureau of Health Manpower 

Notes: Percentages may not sum due to rounding. The number of centers 
reporting in each year starting with 1965 was: 87, 92, 101, 100, 87, 
100, 102, 107, 106, 109. 

1. Professional Fees are incompletely reported. 

2. Defined as indirect cost revenues resulting from sponsored research, 
sponsored teaching and training and other sponsored programs. 

3. Defined as restricted revenues from separately budgeted instructional and 
training programs supported by federal sponsors. 

4. Federal Capitation data as reported by Bureau of Health Manpower and off- 
set by one year to correspond with AAMC Financial Questionnaire. Prior 
to 1971 these data reflect Basic Improvement Grants and Special Project 
Grants; after 1971, Capitation is included. 

5. The "other" category includes a variety of minor revenue sources and it 
is included only to complement the major revenue sources. Fluctuations 
in the "other" category cannot be identified with any particular revenue 



source. 



4-7 



FIGURE 4A-1G 



ACADEMIC MEDICAL CENTER FUNDING, 1964-1974 

(Total revenues for all centers in 
millions of 1964 constant dollars 1 ) 











F 


iscal 


Year 












1965 


1966 


1967 


1968 


1969 


1970 


1971 


1972 


1973 


1974 


Total Operating 
Budget 


751 


804 


881 


961 


906 


1114 


1193 


1290 


1320 


1368 


Federal Research 
Funds 


268 

36% 


280 

35% 


301 
34% 


318 
33% 


259 
29% 


277 
25% 


254 

21% 


281 
22% 


293 
22% 


283 

21% 


State & Local 
Government Funds 


96 
13% 


106 
13% 


124 
14% 


128 
13% 


146 
16% 


168 
15% 


200 
17% 


166 
13% 


239 
18% 


267 
20% 


Professional Fees 
& Service Programs 


29 
4% 


38 

5% 


44 
5% 


71 
8% 


106 
12% 


203 
18% 


234 
20% 


248 
19% 


283 
22% 


296 
23% 


2 
Indirect Costs 


47 
6% 


53 
7% 


52 
6% 


67 
7% 


63 
7% 


69 
6% 


71 
6% 


80 
6% 


91 
7% 


87 
6% 


Federal Teaching 
and Training^ 


92 

12% 


101 
13% 


110 
12% 


125 
13% 


104 
11% 


121 

11% 


132 

11% 


140 
11% 


149 
11% 


143 
10% 


4 

Federal Capitation 


N.A. 


N.A. 


6 

1% 


15 
2% 


20 
2% 


29 
3% 


38 
3% 


49 
4% 


80 
6% 


71 
5% 


Non-Federal 
Research 


42 

6% 


44 
5% 


46 
5% 


45 
5% 


40 
4% 


53 
5% 


58 

5% 


19 
1% 


71 
5% 


25 
2% 


Student Fees 


37 
5% 


37 

5% 


37 

4% 


38 
4% 


33 
4% 


39 

4% 


42 
4% 


50 

4% 


56 

4% 


60 

4% 


Endowments and 
Gifts 


31 

4% 


34 
4% 


33 

4% 


30 

3% 


32 

4% 


34 
3% 


35 
3% 


43 
3% 


43 
3% 


33 
3% 


Other 6 


109 
15% 


111 
14% 


128 
15% 


124 
13% 


103 

11% 


121 

11% 


129 
11% 


214 
17% 


15 

1% 


103 
8% 



Sources: AAMC Financial Questionnaire and Bureau of Health Manpower 

Notes: Percentages may not sum to 100 due to rounding. The number of centers 
reporting in each year starting with 1965 was: 87, 92, 101, 100, 87, 100, 
102, 107, 106, 109. 

1. Professional fees are incompletely reported. 

2. Defined as indirect cost revenues resulting from sponsored research, 
sponsored teaching and training and other sponsored programs. 

3. Defined as restricted revenues from separately budgeted instructional 
and training programs supported by federal sponsors. 

4. Federal Capitation data as reported by Bureau of Health Manpower and off- 
set by one year to correspond with AAMC Financial Questionnaire. Prior 
to 19 71 these data reflect Basic Improvement Grants and Special Project 
Grants; after 19 71, Capitation is included 

5. Deflation performed with Halstead Higher Education Instruction Price 
Index and NIH R s D Price Index. 

6. The "other" category includes a variety of minor revenue sources and 
it is included only to complement the major revenue sources. Fluctu- 
ations in the "other" category cannot be identified with any particu- 
lar revenue source. 



4-8 



FIGURE 4A-1H 
TRENDS IN SELECTED ACADEMIC MEDICAL CENTER FINANCES 
Average of All Centers, 1964—1974 



CURRENT DOLLARS 
(Millions) 



1964 CONSTANT DOLLARS 
(Millions) 



2 0- 



I 5- 



10- 



5- 




-^ 



&**++/ 



I TOTAL 
BUDGET 



r FEDERAL - 
V— RESEARCH 

\^ STATE 
\ FUNDS 



'PROF 
FEES' 



1 1 1 1 :— i n 1 r- r — 

65 68 70 72 74 




2 



■15 



-5 



6s" 1 ' 68 ' 70 ' 72 ' * 



4-9 



centers revenues in current dollars increased 208% to 2.40 billion 
dollars. When adjusted for inflation, this increase was actually 
one of 82% to 1.37 billion constant dollars. Thus, the total bud- 
get revenues of the academic medical centers have increased nearly 
twofold. Inspection of the various revenue categories shows that 
state funds have increased 178%; student fees, 621; endowments and 
gifts, 6%; and reported professional fees increased dramatically 
during the ten year period. During this period also, federal re- 
search funds rose from 268 million constant dollars in 1964 to a 
peak of 318 million constant dollars in 1967. There was then a 
decline followed by the recovery previously noted. By 1973 Fed- 
eral research funds were 9% above 1964 levels. Non-federal re- 
search funds rose slowly to a peak during the period 1969-1971. 
The decline in federal research support during 1969-1971 was also 
offset somewhat by increases in local and state funds which were 
not primarily for the support of research but for the support of 
the service and teaching programs. These state and local funds 
were provided not only to public medical schools but also to most 
privately-owned medical centers. A graphic presentation of the 
total and selected revenue categories averaged for all centers 
reporting is given in 4A-1H in current and constant dollars. In- 
direct costs recovered from sponsored research activities were 
about 6% of total budget and rose roughly parallel to increases 
in total budget. Indirect costs in constant dollars rose 85% over 
the decade as compared to a 6% rise in federal research funds and 
an 82% rise in total operating budgets. 

Figure 4A-1I depicts enrollment and faculty changes. Total 
full-time faculty increased 16,681 to 32,561 full-time faculty 
members in 1974. Three quarters of the increase of faculty members 

4-10 



>- 

H 
,-J 

U 

< 
PL, 

E-h 
2 

PU 

o 

ci 

W 

OS 
H 
H 
2 

W 

u 

< 
u 

t— I 
Q 
W 

is 



w 
Q 
< 
U 
< 



< 

w 

m 
< 



en 



CTl 



to 

CTl 



O 
P. l~- 

nj CTl 
CO i— I 



OS 

o en 
co vO 

•H CTl 
PP rH 



t/1 

p. 
co 

■P 

C 

cu 
u 



Pi 
o 

PL. 



o 

H 



oo 

CTl 



en 
o 



vO 

oo 



|-~ o\° CTl o\ Q *0 o\° CTl oN° 

p-(Nv000rJ-010lO • 

rg ^j- lo , — i i — i lo to «< 

O t-H tH LO 2 

lo eg ,_( to 



*0 o\° OO o\=> i— I o\o \0 o\° *0 o\° 

>*ovoi^oaiLoi^^HO 
lo **/ to t-h \0 lo eg lo 



o o\° lo «\° ^j- o\o oo o\o t>- o\o 
loo^-lolocti^ictiooo 
vo^-loi— to r^ eg *^- 



ai 
to 

■3- 
to 



CO 
LO 
LO 

to 
to 



en 

vO 



tO oN° t-H o\° 
HHOOa 
l/l I s - CO N 



r-- o\q to on° 
■*:$• eg o oo 

O NL/5 N 



CM o\o r^- o\<= 

OHiOOl 
ai t^ LO CNJ 



t^ a\° C^J o\o \0 =N° LO o\« (X] o\° 
COHHinHONOONLO 
^■JOHOHNNN 



t-H o\° tO o\» 
LO O LO O 

^- [^ o to 



CTl o\° *3" o\° t-O o\° ^- o\° KO o\° 

^OH^-Lnaio^ooioo 
O^-i-Hi— IOr— lOrsjt— I 



CS1 



LO o\° r^ o\° 

vO O O O 

^1- r~ eg to 



*o to o\o eg o\« to o\° eg o\o o o\° 
**■ to en ** r» «3- t— ivOvoior~- 

O OOtOOO,— lt-^r-H-3-CN],— I 



to 
eg 



rH 

o 



O o\o OO o\° 
i— I CTl CTl i— I 

en vo o to 



CO o\° OO o\° tO o\° CO o\o LO o\° tO LO o\° OO o\° 

tOOr^-OCTlOcOOOOO vO tOONO 

LO^J-HHOHOINN t-H ^J-t^t^tO 



tO o\° vO o\° ^ o\° O o\° eg o\° 
(NN ini/)\Or- ItO^DLOl^ 
^■^■NHtOHt^fMN 



CD 



eg o\° ^ o\o 
en ai o i— i 
eg \o o to 



LO o\° LO o\«> \0 o\° O o\a *"J" 0\° 
tOtOi-HLOLOCTlLOvOt— It- 
00 •* tO i— I O CTl eg O 



a. 


CH o\° O o\° 


-r 


CO N^O fO 


i— l 


^- \0^0 m 



O OO oN=> 00 o\° CTl o\° i— I o\° 

o csioor^voooocnoo • 



r—1 a\° K) oNO 

oo r- K) Kl 





■IH 


■M 


c 








d 


(=: 


cu 






■p 


CO 


cu 


f= 






c 


e 


t-H 


t-H 






cu 


t-H 


03 


rH 






6 


T-H 


> 


o 


1/1 




t-H 


o 


•H 


in 


-(-> 




r-H 


!H 


D 


C 


c 




o 


C 


cr 


w 


CU 




(I 


P-l 


w 




TJ 




c 






■p 


■H 




w 


+-" 


■M 


c 


[/) 






c 


c 


cu 


CU 


CO 


+-1 


CU 


cu 


-d 


OS 


t-H 


c 


t3 


T3 


D 




nj 


0) 


D 


D 


4-> 


T3 


JH 


-n 


4H 


+-> 


CO 


c 


o 


3 


CO 


co 




aj 


4-> 


+-> 






cu 




u 


to 


t-H 


r-H 


4H 


tfl 


o 




aj 


aj 


CO 


C 


a 


H 


U 


U 


3 


!H 


r 


a) 


•tH 


•H 


-d 


CU 


4-> 


+h 


13 


TT 


nj 


4-J 


CO 


O 


CU 


CO 


!-l 


fS 


o 


H 


t; 


^ 


u 


1— 1 


Cu 





>. 




>. 


+H 




+J 


t-H 




t-H 


D 


>, 


~ 


u 


*-> 


CJ 


aj 


tH 


m 


PP 


D 


PL, 




CJ 




cu 


nj 


C) 


o 


Li, 


e 


C 




■H 


cu 


Co 


H 


•H 


O 


1 


<J 


C 


t-H 


CO 


CO 


t-H 




•H 


p 


T—t 


U 


PL, 


03 
U 


co 


t-H 


• H 


o 


03 


C 


■H 


+H 


•H 


CO 


o 


t-H 


CTJ 


H 


U 


m 






CO I 

p. ^1- 

•H vO 

aj en 

C t-H 

c 
o 

■h ; 

■M <S1 

CO CO 

CO 4H 

d nj 
CO 



nj 
O 

PJ 



o 



= 

6 
o 
c_> 

c 
o 

CO 



e 

CO 



p^ 



3 



aj 
O 
• H 

CU 

S 

"3 
CU 
+J 

o 

3 
p. 
»-• 
co 
C 



P. 
nj 
Cu 
>, 

C 

•H 

-C3 
CO 
CO 

u 
u 

3 



-d -d 

c 

-t-i 03 

C 

CO vO 

e ^o 

t-H CTl 

i— I t-H 

o 

SH U 

C O 

<D <+-| 

LO aj 

vO +-1 

CTl 03 

,-H Q 

H CN1 



O 

u 

c 

CO 

■rH 

CO 



n3 
CO 

^: 

p. 
co 



c 

CO 



u ■ 
e co -d 
CO X) cu 

■-H 6 +-> 

P. 3 P. 

CP C O 



a. 

rH CU 
03 P. 

o 

•H >, 
■P t-H 
CO X) 
(H CO 
O -H 
CO P. 
X nl 
H > 
to 



4-11 



were in the clinical sciences which comprise two-thirds of the 
total faculty manpower in U.S. academic medical centers. In the 
same period of time, basic science faculty increased by 4,300 per- 
sons to a total of 9,500 basic science faculty members in 1974. 
The ratio of clinical to basic science faculty increased slightly 
from 2.04 to 2.45. From 1970 to 1974 part-time faculty declined 
slightly but clinical volunteers increased by 12,000 persons. 

The burgeoning health science student enrollment during the 
last ten years is also shown in Table 4A-II. Medical student en- 
rollment increased by 541 from 32,100 total medical students in 
1964 to 49,600 medical students by 1974. Graduate student enroll- 
ment doubled from 5,500 to approximately 11,000 students and post- 
doctorals also doubled although the number (2,450 by 1974) was much 
smaller. House staff training in the academic medical centers in- 
creased by 15,000 persons from 19,000 in 1964 to 34,400 ten years 
later. Finally, medical student equivalents increased by 57% 
(6,100 students) to a total of 16,800 student equivalents in 1974. 
("Medical student equivalents" are defined by the Association of 
American Medical Colleges according to a formula which includes 
dental, nursing, pharmacy and other allied health students depend- 
ing on the intensity of educational effort provided by the medical 
school faculty.) 

3 . Established Academic Medical Centers (N=86 ) 
Part of the increases in funding, enrollment and faculty dur- 
ing the period 1964-1974 are due to the founding of approximately 
30 new medical schools and not entirely to real growth of existing 
schools. In order to separate the expansion of previously existing 
schools from the increment due to new schools, this section pre- 
sents data for the 86 academic medical centers which existed at 



4-12 



the beginning of the period 1964-1974. Exhibits 4A-2A through 
4A-2E display the data for the 86 established academic medical 
centers . 

In constant dollar terms, the absolute amount of federal re- 
search funds increased slightly between 1964 and 1974. However, 
federal sponsored research funds declined from 361 of the total 
revenues of the academic medical centers in 1964 to 23% in fiscal 
1974. An increase in the emphasis of medical centers on a wider 
variety of programs is seen when one considers state and local 
government funds and the sum of professional fees plus service 
programs. The latter two categories are directly linked to patient 
care and community service. In fiscal 1965, these service cate- 
gories totalled 101 of the academic medical centers revenues; by 
fiscal 1974 they had risen to 26%. There was also a significant 
rise - particularly in the publicly-owned medical centers - in 
state and local government funds which are provided primarily for 
the direct support of teaching and of faculty. Of course, these 
funds indirectly support research by supporting faculty who engage 
in federal and non-government sponsored research activities. 

In other words, in 1964, the federal research effort carried 
out by academic medical centers was the largest single component 
of the academic medical centers' revenues. Ten years later, al- 
though the constant dollar volume of federal research activity was 
approximately the same, the service and teaching activities of the 
academic medical centers had increased so as to equal those of the 
federal research activity as measured in absolute constant dollar 
terms. The average medical center thus appears in 1974 to be more 
diversified and to be less dependent on a single source of revenue 
for its essential activities of reasearch, service and teaching. 



4-13 



Tables 4A-2C and 4A-2D should be compared with Tables 4A-1E 
and 4A-II (comparable tables for all centers) and with comparable 
tables in subsequent groups. Table 4A-2D shows the faculty/enroll- 
ment changes for the 86 medical schools which were established in 
1958. Although approximately 30 new schools were started during 
the decade 1964-1974, the 86 established medical schools enlarged 
their faculties by 87 percent during the same period. The increas- 
es in students previously commented upon are again noted. In 
Figure 4A-2E selected categories of student and faculty data are 
displayed graphically. Teaching load of basic science faculty is 
also shown in this Figure and subsequent graphs. 

4. Public Medical Centers (N=43 Exhibits 4A-3A to 4A-3E ) 

Public institutions had a rapid rise in total budget. Aver- 
age total operating budget increased approximately 75% from 8 mil- 
lion to 14.2 million constant dollars. Student fees in public in- 
stitutions accounted for about 3% of budget revenues and state and 
local funds were about 30%. In absolute dollar amounts state and 
local funds increased from 2.2 to 4.5 million dollars. Public 
medical centers were lower in the endowment and gift category as 
compared with private centers. Endowment and gifts ranged between 
1 and 2%. Federal research funds were slightly below but closely 
parallel to those of the average of 86 centers falling from 321 to 
20% of total budget over the ten years. Research funds rose slight 
ly in constant dollars. The total of service programs plus profes- 
sional fees rose from 8% of revenues in 1965 to 21% in 1974. 

Public centers tended to have fewer faculty than the average 
of 86 centers and to have more students, particularly undergraduate 
medical students and medical student equivalents. Therefore, basic 
science faculty teaching load is higher at public institutions. 

4-14 



5. Private Medical Centers (N=43, Exhibits 4A-4A through 
4A-4E 

Average total operating budgets for the privately-owned cen- 
ters were approximately the same as for the average of all centers 
and publicly-owned centers, rising from 9 million to 15 million 
dollars over the decade. Federal research funds were higher than 
the average of all centers and about 1 million dollars higher 
than public center research funds at comparable times. Research 
funds increased from 3.7 million constant dollars in 1965 to 4.0 
million in 1974. State funds provided to private centers were 
fairly stable ranging from 3 to 5 percent of total budget over 
the decade. A much smaller percentage than in public medical cen- 
ters (27-321). Endowments and gifts varied from 8% to just under 
6%. Professional fees plus service programs comprised about 30% 
of the total budget by the end of the decade up from 121 at the 
beginning. Faculty size was considerably larger than average, 
student body size smaller and, consequently, teaching load was 
lighter. 

6 . Centers Ranked by Research Involvement (Exhibits 4A-5A 
through 4A-8D ) 

Since this study focuses on federal biomedical research fund- 
ing in academic medical centers, an effort has been made to devise 
a scale of federal biomedical research involvement. The details 
of the procedure are described in the Methods Section. The ef- 
fects of size have been factored out and the research involvement 
scale attempts to integrate a number of variables related to the 
intensity biomedical research activity at each center. The 86 
established medical centers were ranked accordingly to this scale 
of research involvement and divided into research quartiles for 
comparative analysis. 

4-15 



Centers ranking in the first quartile for research involve- 
ment had total expenditures averaging 60 to 701 above the mean 
for 86 established institutions. Federal research funds received 
in 1964 were 43% of total budget (86 center average = 36%) and 
declined to 32% (average 241) by 1975. However, in constant dol- 
lars federal research revenues actually rose from an average of 
$6.3 million in fiscal 1965 to $7.6 million in 1974. As in other 
centers, service and professional fees rose to about 25% of reve- 
nues by 1974. State and local funds averaged 7-8% in first re- 
search quartile centers due to the preponderance of private insti- 
tutions in this quartile. Because federal research revenues were 
greater in centers in the first research quartile the sum of reve- 
nues from all federal sources (research, teaching and training, 
special project and capitation grants) was 54% of total budget as 
compared to an average of all centers of 45%. 

Centers in the lowest quartile had lower total budgets, lower 
total federal funds and federal research funds and much more state 
funds. Federal research funds were 32% of total revenues in 1965 
but fell to 11% in 1974. State revenues and service income were 
each much greater than research funds in this group. 

Examination of student/faculty patterns in these research- 
ranked centers shows that f irst-quartile centers had the largest 
faculties but also the largest number of total students. There 
were also 38% more graduate students in research intensive centers 
than in the average center in 1974. Conversely, the lowest number 
of total students, graduate students and medical student-equivalents 
(but also faculty members) were in the least research intensive 



4-16 



centers. Average medical student enrollment and teaching 

load was about the same in all quartiles. 

7. Geographic Regional Differences (Exhibits 4A-9A through 
4A-12D ) 

In comparing the academic centers by geographical regions 
(Northeast, South, Midwest, West) the most striking differences 
are in federal sponsored research funds as a fraction of total 
revenues. Federal research activity in western and northeast cen- 
ters was above the average for all centers in federal research ac- 
tivity throughout the decade. The western centers particularly 
maintained a high ratio of federal research funds to total budget 
being 411 at the beginning of the decade and 321 at the end of the 
decade. Northeast centers declined somewhat more but were still 
above the average. In contrast, southern centers were consistent- 
ly below the average in federal research funds. Research revenues 
in southern centers fell from 321 in 1965 to 191 in 1974. South- 
ern centers were considerably above average in receipt of state 
support (241 in 1974) and in professional fee income (reporting a 
high of 17.5% in 1974). Midwestern institutions were also below 
average in federal research revenues throughout the decade, re- 
ceived a relatively high percentage of state funds and had about 
average professional fee income. Northeast centers had the lowest 
reported professional fee earnings [1% in 1974) followed by mid- 
west (11.3%) and south (17.51). The growth of total budget expen- 
ditures was highest in the west and northeast (from 10 million to 
about 17 million), lowest in the southern centers (an increase 
from 6.6 to 12.1 million) and intermediate in the midwest (growth 
from 8.6 to 14 million constant dollars). 

Total student enrollment was greatest in the western centers 



4-17 



and lowest in the south. Medical student enrollment was lowest 
in the west and highest in the midwest. Midwestern schools had 
the largest graduate student populations and the heaviest teach- 
ing loads. Northeastern and southern schools enrolled the small- 
est number of other health science students (medical student equiv- 
alents) and also had the lowest student/faculty ratios. 

8 . New Medical Centers 

Approximately 30 new medical schools came into being at var- 
ious times between 1964 and 1974. This continuously expanding co- 
hort of schools makes it extremely difficult to chart the under- 
lying trends in finance, enrollment and faculty associated with 
the new schools. In an effort to separate the trends of federal 
research funding in developing schools from changes due to accre- 
tion of additional schools, we have chosen a special method of pre- 
sentation for the new schools. This method consists essentially 
of synchronizing the time profiles of the new schools by year of 
operation rather than by calendar year*. In the application of 
this method, there is a trade-off between the number of schools 
and the number of years included in the data set. The number of 
school-years (analogous to man-years) in the sample could be maxi- 
mized by tracking ten schools over a six-year time interval. 

Exhibits 4A-13A through 13D display the funding, enrollment 
and faculty size of ten schools as functions of year of operation. 
Federal research funds in constant dollars more than doubled dur- 
ing the first six years of operation, but declined from 30% to 
221 as a proportion of total operating budget. As in the case of 



*That is, year one is the first year in which the school 
enrolled a class of medical students, etc. 



4-11 



established schools, service programs and professional fees in- 
creased from 6% to 25% of the total operating budget. The most 
distinctive feature of the new schools was their heavy dependence 
on state and local government funds, averaging over 30% for the 
six-year time span. In nearly all aspects, except state and 
local government support, the funding pattern of the new schools 
was similar to that of the established centers. As expected, the 
new schools achieved the same proportional enrollment of medical 
students as the established centers within four years after open- 
ing and the proportions of house staff and medical student equiva- 
lents were similar to those of established centers. After six 
years of operation, the new school tended to have proportionally 
more graduate students and fewer post-doctorals than the estab- 
lished centers. 



4-19 



TABLES AND GRAPHS 
FOR 
GROUPS OF ACADEMIC 
MEDICAL CENTERS 



4-21 



Table 4A-2A 
FUNDING IN ESTABLISHED ACADEMIC MEDICAL CENTERS 
1964-1974 
Average for 86 Medical Centers in Thousands of 1964 Constant Dollars 

Fiscal Year 
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 



Total Operating Budget 


8766 


9381 


10068 


11048 


11594 


12467 


13294 


13788 


14794 


14751 


Federal Research Funds 


3194 2 


3370 


3577 


3816 


3539 


3327 


3064 


3440 


3686 


3456 




36% 


36% 


36% 


35% 


31% 


27% 


23% 


25% 


25% 


23% 


State and Local Govern- 


1269 


1367 


1534 


1633 


1664 


1945 


2271 


2008 


2169 


2561 


ment Funds 


14% 


15% 


15% 


15% 


14% 


16% 


17% 


15% 


15% 


17% 


Professional Fees and 


919 


1086 


1161 


1569 


1722 


2620 


3073 


3208 


3187 


3785 


Service Programs 


10% 


12% 


12% 


14% 


15% 


21% 


23% 


23% 


22% 


26% 


Indirect Costs 4 


552 


628 


632 


782 


838 


831 


857 


947 


1070 


1029 




6% 


7% 


6% 


7% 


7% 


7% 


6% 


7% 


7% 


7% 


Federal Teaching and 


1059 


1181 


1262 


1453 


1380 


1384 


1524 


1600 


1730 


1606 


Training^ 


12% 


13% 


13% 


13% 


12% 


11% 


11% 


12% 


12% 


11% 


Federal Capitation 


N.A. 


N.A. 


68 


182 


370 


438 


503 


697 


964 


866 








1% 


2% 


3% 


4% 


4% 


5% 


7% 


6% 


Non-Federal Research 


962 


1016 


1042 


1059 


1206 


1264 


1339 


1306 


899 


899 




11% 


11% 


10% 


10% 


10% 


10% 


10% 


9% 


6% 


6% 


Student Fees 


428 


438 


441 


455 


452 


475 


516 


607 


670 


720 




5% 


5% 


4% 


4% 


4% 


4% 


4% 


4% 


5% 


5% 


Endowments and Gifts 


623 


649 


676 


684 


771 


732 


761 


889 


514 


624 




7% 


7% 


7% 


6% 


7% 


6% 


6% 


6% 


3% 


4% 



Sources: AAMC Financial Questionnaire and Bureau of Health Manpower 



Constant dollars computed with Halstead Higher Education Price Index and NIH R & D 
Price Index. 

Percentages and dollars may not sum to total due to averaging, rounding and incomplete 
reporting of data. 

Professional fees are incompletely reported. 

4 
Defined as indirect cost revenues resulting from sponsored research, sponsored teaching 
and training and other sponsored programs. 

5 
Defined as restricted revenues from separately budgeted instructional and training 

programs supported bv federal sponsors. 

Federal Capitation data as reported by Bureau of Health Manpower and offset by one year 
to correspond with AAMC Financial Questionnaire. Prior to 1971 these data reflect Basic 
Improvement Grants and Special Project Grants; after 1971, Capitation is included. 



4-22 



FIGURE 4A-2B 



AVERAGE MEDICAL CENTER BUDGET 



Average of 86 Centers, 1964-1974 



Percent of 
Total 1964 Budget 



F^rcent of 
Total 1974 Budget 



Trends in Selected Categories 
Millions of 1964 Constant Dollars 



Other 




-15 



i**** 



Total Budget j 



,""' 



y 



-10 



rJ' 



Research 




State 



Pro. Fees 



Other 



Pro. 
Fees 



State 



Fed. 

Resj 



65 66 67 68 69 70 71 72 73 74 



I 



26% 



17% 



23% 



Based on Table 4A-2A 



4-23 



TABLE 4A-2C 

BIOMEDICAL AND BEHAVIORAL RESEARCH FUNDING 
IN 86 ESTABLISHED MEDICAL CENTERS 1 



Averages in Thousands of 1964 Constant Dollars ' 



Program 



1968 



1970 



1972 



1974 



Regular Research 
Grants 



$1,959 3 $1,648 $1,879 $2,214 



Program Project and 
Center Grants 



613 



615 



1,042 



1,259 



Contracts 



18 



927 



Training Grants and 
Fellowships 



1,214 



1,103 



1,036 



1,086 



Total from NIH 



5,377 



3,538 



4,278 



5,956 



Total from NIMH' 



525 



494 



402 



407 



"Data from NIH/IMPAC file. 



Constant dollars computed using NIH R&D deflator. 

Program totals will not sum to NIH total because only selected 
programs are shown. 

Excluding service programs. 



4-25 



Total Student Enrollment 
Medical Student Enrollment 

Medical Student Equivalents 

Graduate Student Enrollment 

Interns and Residents 

Post-Doctorals , Basic Science 



TABLE 4A-2D 

ACADEMIC MEDICAL CENTER ENROLLMENT AND FACULTY 

1965-1974 

Average Values for 86 Schools 

1965 1966 1967 1968 1969 1970 1971 1972 1973 



1974 



792 


833 


849 


884 


954 


990 


1005 


1100 


1205 


1238 


370 


375 


381 


388 


392 


416 


440 


464 


499 


524 


47% 


45% 


45% 


44% 


41% 


42% 


44% 


4 2% 


41% 


42% 


123 


130 


135 


146 


140 


159 


168 


202 


216 


222 


16% 


16% 


16% 


17% 


15% 


16% 


17% 


18% 


18% 


18% 


63 


81 


83 


81 


98 


94 


99 


104 


108 


120 


8% 


10% 


10% 


9% 


10% 


10% 


10% 


10% 


9% 


10% 


219 


230 


233 


252 


282 


309 


300 


332 


354 


376 


28% 


28% 


28% 


29% 


30% 


31% 


30% 


30% 


29% 


30% 


15 


15 


14 


15 


52 


35 


19 


27 


32 


25 


2% 


2% 


2% 


2% 


6% 


4% 


2% 


2% 


3% 


2% 



Total Full-time Faculty 
Clinical Science Faculty 

Basic Science Faculty 



L82 



196 



220 



250 



251 



261 



285 



316 



344 



340 



122 


131 


153 


176 


177 


186 


200 


228 


250 


243 


67% 


67% 


70% 


71% 


71% 


71% 


70% 


7 2% 


73% 


71% 


60 


65 


67 


73 


75 


77 


84 


87 


94 


95 


33% 


33% 


30% 


29% 


29% 


29% 


30% 


28% 


27% 


29% 



Numbers may not sum due to rounding and averaging. 

Theoretical number of other health science students instructed by medical school faculty. 



4-26 



FIGURE 4A-2E 



ENROLLMENT AND FACULTY 



Teaching Load* 

Medical Students- 
Total Students 

Faculty — 1«*' 



Relative Changes in 
Students and Faculty 

1972= 10 



1964-1974 




2 



65 66 67 68 69 70 71 72 73 74 



Percent of Total Students 
by Categories 




100% 



POST -DOCTORAL 



EQUIVALENTS 



80% 



60% 




40% 



20% 



£1 0% 



65 66 67 68 69 70 71 72 73 74 



Average for 86 Schools 



•TEACHING LOAD = (1/2 MEDICAL STUDENTS AND 1/2 MEDICAL STUDENT EQUIVALENTS AND GRADUATE STUDENTS AND POSTDDCTDRALS )/ 
BASIC SCIENCE FACULTY. 



tINDEX NUMBERS - SEE METHODS SECTION. 



ACADEMIC MEDICAL CENTER GROUPS 

EXHIBITS 4A-3A THROUGH 4A-13D 
EXPLANATORY NOTES 

AMOUNTS SHOWN IN FINANCIAL TABLES ARE IN THOUSANDS OF 1964 
CONSTANT DOLLARS. THE HALSTEAD HIGHER EDUCATION PRICE INDEX 
AND THE NIH R&D PRICE INDEX WERE USED FOR DEFLATION. AN 
ASTERISK (*) IN THE TABLE INDICATES AMOUNTS LESS THAN $500. 



PERCENTAGES SHOWN IN FINANCIAL TABLES REPRESENT PERCENT OF 
TOTAL OPERATING BUDGET. THE REVENUE SOURCES SHOWN IN THESE 
TABLES DO NOT ACCOUNT FOR THE ENTIRE OPERATING BUDGET. 



ALL DATA EXCEPT FEDERAL CAPITATION AND SPECIAL PROJECT GRANTS 
WERE OBTAINED FROM THE ANNUAL QUESTIONNAIRE OF THE LIAISON 
COMMITTEE ON MEDICAL EDUCATION. 



FEDERAL CAPITATION AND SPECIAL PROJECT GRANTS DATA WERE OBTAINED 
FROM THE BUREAU OF HEALTH MANPOWER AND REPRESENT FUNDS OBLIGATED 
IN A GIVEN YEAR RATHER THAN FUNDS EXPENDED BY THE MEDICAL CENTER. 



PERCENTAGES SHOWN IN THE ENROLLMENT/FACULTY TABLES REPRESENT 
PERCENT OF TOTAL STUDENT ENROLLMENT AND PERCENT OF TOTAL FULL- 
TIME FACULTY, RESPECTIVELY. 



THE TERM "MEDICAL STUDENT EQUIVALENTS" DESIGNATES THE THEORE- 
TICAL NUMBER OF OTHER HEALTH SCIENCE STUDENTS INSTRUCTED BY 
MEDICAL FACULTY. 



7. THESE TABLES REPRESENT VARIOUS GROUPS OF ACADEMIC MEDICAL CEN- 
TERS. DUE TO INCOMPLETE REPORTING OF VARIOUS FINANCIAL AND 
ENROLLMENT DATA, THE NUMBERS MAY NOT SUM TO THE COLUMN TOTALS. 

8. TO FACILITATE COMPARISON OF CENTERS AND GROUPS OF CENTERS 
WITHOUT REGARD TO SIZE, THE FINANCIAL, ENROLLMENT AND FACULTY 
DATA ARE REPRESENTED GRAPHICALLY AS INDEX NUMBERS . THE INDEX 
NUMBER REPRESENTS THE VALUE OF A VARIABLE IN A GIVEN YEAR 
RELATIVE TO A VALUE OF 1.00 IN AN ARBITRARILY CHOSEN REFERENCE 
YEAR. THE REFERENCE YEAR FOR THE FOLLOWING GRAPHS IS 1972. 
SEE METHODS SECTION FOR DETAILS. 

9. IN THE FINANCIAL GRAPHS, FEDERAL RESEARCH REVENUES, STATE AND 
LOCAL GOVERNMENT FUNDS AND SERVICE PROGRAMS AND PROFESSIONAL 
FEES ARE INDEXED ON THE 1972 VALUE OF TOTAL OPERATING BUDGET 
RATHER THAN ON THEIR OWN VALUES IN 1972. 



4-29 



< 



\] OJ O 
\l * O 


on x 
men 


on x 

W 00 


en x 
en oj 
in tr» 




o 

OJ 


nO X 

o * 


0\ X 


in x 
r- in 
so r^ 


4" X 

or- 

* OO 


CO On 


in ao 
njrv 


^ x 

On OJ 

-* in 


* m 


-*■ 


--« «H 






»-« 


OJ O 


o 


-* 


"-* (T\ 


in 


*■* 


-* O 


m r- x 
\) on en 

i^ mco 


iri x 

On 'O 

nO o 


00 X 
00 O 

no m 


OJ x 

r^oo 




X 


r^ x 

on -* 


O X 

--• o 

*-* 00 


w x 

in ■*- 
so r- 


*c X 

00 OJ 
* 00 


OJ x 

in co 


in x 

* 00 

ojr^ 


m x 

On ^ 
O On 


m oj 

r-* 


in 


— > OJ 


oi 




o 


Oj *"l 

OJ 


o 


4- 


-H O 


in 


rH 


■h r. 


~4 m x 
n so oj 


in X 

00 O 

it r* 


m x 


-< x 


OJ 
O. 


X 

* 


■*- x 

00 CT> 

•c in 


o X 

•-4 <■-< 


oj in 

r- oo 


■* X 

«* Ol 


r- x 

O CO 


m x 

-* 00 

r-on 


-H X 

in r- 


M OJ 


* 


*-• »-• 


o> r- 




'VI 


OJ *-f 


r-. 


in 


«-• On 


in 


in 


T-* T-« 



I-- X 


< « 


00 X 


* « 


m 


j* 


ce ~* 


»- o 


O <C 


■< * 


m 


m 


OJfO 


< o 


* ro 


o-nn 


m 


*o 



i^ oo 






00 f 



mo 



04 S* 

00 4 

oj m 
^ o 



O r. .-. 



O) »< 
oo O 

oj in 



■^ x 


OJ X 


O X 


■rH X 


o« on 


o oi 


in a\ 


oo r-- 


oj m 


^ o 


m o 


4T1 -O 



in .-e 


r- x 


* « 


00^ 


in ** 


%o in 


00 Ol 


--! OJ 


in oj 


* m 


r- c^ 


o in 


•O * 


in a\ 


o OJ 



o ^^ 



m in 

oj >o 



OJ ?* 


m •■< 


0^ *< 


o « 


* 


»» 


O i» 


m « 


w»< 


m « 


m * 


OO OJ 


oo in 


•o in 


01 


OJ 


•o o-i 


O 00 


* ^ 


m t^ 


ow 


OJ no 


<D Cl 


•< * 


«j 


^1* 


v* OT 


r^ -m 


oa«« 


a\ * 



* » 


< « 


in >* 


O c* 


CO OJ 


t- o 


m r^ 


■rt Ol 


OJ 00 


•< o 


ro oj 


oo <n 



O J* 


00 *< 


en « 


r^ * 


in « 


m o> 


rv m 


■4) t^ 


OJ CO 


f^m 


■rf M 


r< rs 


oi -a 


-O OJ 


nc 



oo r- 
o. c 



4 M 


C 1 


a *e 


■H X 


m: «* 


ir m 


0" w 


* m 


o- « 


>c in 


o *-» 


<o o 


* 0" 


on >c 


OJ Ol 


OD r- 


* in 


00 OC 


•« •- 


o. •- 


o- o- 


■r- 4 


■f o- 


*- rr 


^» -- 


ir t- 


•f o. 



< 

< 



UJ 

< 
a. 

UJ 

> 



ir 

oo 



NO W 


o « 


00 ^» 


OJ *o 


c •* 


dj •- 



in x 
on i^ 



o\ *< 


OJ « 


* X 


c^ ~o 


o in 


mm 


--* 


if. a 


-» o 



OJ .-< 


-< x 


en x 


cox 


i) 


»•* 


O X 


en X 


co on 


f- o 


oo r^- 


on th 


r* 


U 1 


O 00 


.-t o 


OJ * 


< o 


en — . 


«-< OJ 


--* 


rl 


sO -r* 


^ * 


• 


o • 


• 


•* • 




• 


•» ° 


• 


m 


o 


OJ 


oj r- 




OJ 


OJ OJ 


--* 



on x 
*o m 



*€ 


O J* 


-o « 


() 


on on 


o o 


in 


* m 


<i m 



■^ OJ 









< 
(1 

n 

<D 

or 



ID 

Q 






o 

3 



■< — . 






z 



a. 

en 

_l 



U I- 



z 



3 



> •-. 



z 



4-30 



xasana iv±oi =10 ±N3^d3d 



CJ3 

5 










s 


\x 


*■ 


\ V 






k. 


' A 


^ 


\A 


*- 


1 \ 


«, 
«. 


K \ 


« 
«. 

^ 
^ 

■*., 


H 


V 


i ! 



dBQWIN X30NI 

4-31 



TABLE 4A - 3C 

BIOMEDICAL AND BEHAVIORAL RESEARCH FUNDING IN 43 PUBLIC MEDICAL CENTERS* 
Averages in thousands of 1964 constant dollars; NIH R&D Deflator 



1968 



Regular Research Grants 1601 



1970 



1386 



1972 



1540 



1974 



1886 



Program Project and 
Center Grants 



604 



577 



946 



1106 



Contracts 



26 



923 



Training Grants and 
Fel lowships 



1084 



955 



886 



925 



Total from NIH 



3859 



3181 



3649 



4875 



Total from NIMH 



418 



396 



313 



325 



*Data from NIH/IMPAC File 



4-33 



rw 


cu x 


in x 


no $* 


00 tf< 


-o *e 


r^ 


in n 


•-«o 


m t* 


r^ «o 


OJ ^- 


m 


in oj 


m o 


*-» r* 


n in 


en 



<J\ X 


in ** 


*-* *o 


en rn 


ai ■& 


m 



r^ x 


en *« 


cu « 


r- a* 


oj m 


(T> CO 


aj n 


-* CO 


in r- 


rvjin 


-t OJ 


ro -r-. 



<t if 


O" X 


r* OJ 


oa r^ 


(VJ CO 


•H 



ai 


en 


m 


■h* 


ai x 


*-• X 


--• ** 


m j* 


r* 


o 


a> 


o 


00 ■* 


-i o 


m m 


OJ OJ 


D 


a 


j 


00 


ai m 


«H AJ 


or -* 


en 


^H 
















»^ 


T « 




o 


m 


o\ 


r-* 


H 


01 






* 


oj 




(V! 




.-« 
















■^t 


DO 


00 


<*• 


ox 


OJ X 


o »< 


00 X 


r- 


a 


■a 


wH 


■* * 


-H O 


a> o 


■wi m 


> 


.-» 


^*- 


m 


fVJfv 


»-* *-* 


OJ OJ 


41 


o 


^ 




« 


• 


• 


• 


• 


r*. 


•H 




OJ 


■H 


O 


41 


*-* 



o m 

oj in 



o *< 


-r-t X 


o 


r- 


en xi 


a0 00 


r« 


■wt 


^ en 


o 




m 



a 


r- 


-* 


=*? 


ca X 


en x 


OJ X 


m x 


r^ 


en 


-*■ 


fXl 


r-*(T\ 


O sO 


en ^o 


n m 


■ 


o 


j- 


Lfr 


OJ 00 


T-* cn 


OJ nO 


o 


O' 


•, 




a 


• 


• 


• 


* 


x 


~-< 




o 


en 


en 


vC 


m 



in x 


-* x 


r- oj 


rv * 


*-* r* 


pi 



z 

UJ 

o 



OJ *< 


OJ X 


o x 


en x 


t-* rv 


00 OJ 


--• o 


en en 


* ■»* 


•**cn 


t* m 


OJO*) 



x) ** 


OJ w 


so m 


r- en 


r4 CX) 


a 



Q 

UJ 

5: 
o 



OJ X 


in x 


o *? 


OJ X 


OJ ** 


-1 * 


en en 


en th 


in oo 


--HO 


-* o 


--*m 


-* 


OJ OJ 


ro 



* C J« 


o_ »♦ 


o. w 


o X 


■X CO X5 


en ^» 


OJ OJ 


»-» -* 


[V -- oc 


o. 


O. If 


*" 



o X 


,-t X 


-1 


r^ oo 


r- -H 


»- 


--• ■* 


10 


<* 


• 


• 


C! 


o 


m 




r* 


a 




c * 


^» «< 


<1 


* oo 


sfl m-i 


»- 


— X 


f 


t 



E 


LU 


LU 


O 


Q 


< 


< 


a 


O 


UJ 


< 


> 



o 


M 


(\J i* 


DO .-< 


* « 


CT> 


mi 


00 ^0 


00 O 


OJ o 


rr 


-£ 


-- m 


■t 


Ou -- 



in 


tn 


-» >« 


-o *e 


^ *< 


OJ *< 


sC 


r»j 


oo in 


sO -< 


r- in 


o en 





on 


en 00 


--» 00 


00 


OJ o 


-* 




• 


• 


o 


• 


* 




m 


en 


CO 


^» 


ir. 




* 


^* 




(VJ 



CO X 

inm 



* X 


^t X 


„-. en 


in o 


w-t 00 


<H 



) 



> 



CJ 



Q 

3 






in 



3 _l 



4-34 



siNaanis nvioi do iNBDusd 




(J> l-H 


ul 


LU I/) 


^ t/> LJJ 


H 


JHh 


>- < O 


2: 


< < 2: 


imz 


UJ 


O => LD 


O LU 





>-h O O 


< tr .-. 


Z5 


O < => 


tuou 


h- 


LU CE |- 




aaawnN xboni 
4-35 



r- « 


oj *< 


.-• *« 


o « 


en, « 


o « 


OJ « 


ooo 


w-i *o 


Ol o 


vO *H 


O 30 


in .0 


o o 


o in 


'H OJ 


in * 


•O fl 


oo >o 


o * 


* 



«-t 


*e 


■* *c 


rv 


O 


oj en 


OJ 


m 


O sO 



in j* 


OC J^ 


oj *« 


^ £* 


m ** 


o ** 


in *j 


m 2* 


0^ w 


O) <J\ 


O -XI 


r^ ^o 


in o 


en ro 


CO 0J 


vO -"« 


i*- * 


m >o 


CO 00 


r^ * 


r- *-* 


\0 r* 


OO o 


•* O. 


-* 


o> »~« 


r- oj 



o « 


O K 


oj n 


r* * 


o -♦ 


O l~* 



UJ 

-I 
en 
< 



^ «c ^t 



-* -H *-l 



in »* 


in *? 


r- cr 


•a m 


in r* 


r- *c 



CO w 


r*- rf< 


in * 


O i* 


*-< *e 


rv >« 


a r- 


r^ fi 


en oj 


s0 X) 


in *o 


in o 


oj m 


r** m 


m oc 


.-. vC 


OJ O: 


m en 



»-• nj oo 



%o <* 


< X 


OJ *< 


in >? 


* « 


m oo 


1- o 


* in 


oo oj 


ii ao 


r^ ^-i 


■* o 


•o in 


•43 oo 


oo O 



in *« 


o *e 


j- s< 


•X) ** 


ro *e 


oo in 


co -t 


oo n 


ro o 


r» in 


■* -*• 


in ao 


O vO 


^ T-l 


■^ OJ 



O 


in 


< > 


« 


«* 


x 


OJ 


rf? 


^» 


*? 


o X 


O 


<.-# 


OJ 


»■* 


CO 


M 


|s 


M 


ro 


x 


rv 


.-e 


in 


^ 


r> 


•o 


tv 


o 


»- 


o 


<T> 


+ 


* 


•in 


m 


r- 


m 


o 


m 


o 


O s0 


00 


<■> 


* 


n 


OJ 


T-* 


cm 


•^i 


■ 


■o 


<J 


en 


< 


o 


* 


<o 


in 


en 


<<j 


o 


a> 


ou 


«-■ 


~* 


^*- 


ou 


-■* 


o 


o 


NO 


f) 


r^ 


en 


r^ 


>jj 


n 




* 


o 


.H 


o 




ei 




4. 


m 


rr 




,. 


«-t 


o 


OJ 


^ 


.< 


r 


^ 


cr 




•* 





o: 


L3 


UJ 


Z 


1- 


i — i 


z 


D 


LU 


z: 


u 


o 




u_ 


J 




< 


a: 


o 


LU 


w 


1 


a 



0J »< 


-* 


«^ 


1^ 3< 


in *< 


OJ « 


in *? 


oo >? 


^f ro 


K- 


c> 


■^ * 


in o~» 


<J\ nO 


in ro 


r^ in 


>£ OJ 


< 


o 


-* LA 


en no 


O0 OJ 


o o 


«« -* 



•O IM 



O « 

o> o 



00 « 
0^ oo 



O »H ^ 



n -^0^ 



OC H rH 



< o 



♦ •-! P^ if 



in m 



in en 
rvj o 



o « 



■r-irv 

m «o 



-« in 
in 



< 



r 



r- ffi -< 






OC ** 


cr. a* 


•- *« 


■« cr» 


r- oj 


--< OJ 


r- «- 


o cr 


< r- 



IV «- ^ 



< 
< 



in 
UJ 
13 
< 
O. 
UJ 

> 
< 



in 



a 
m 



en oo 
it a- 



in ^ 


OJ i^ 


m ■* 


OJ i* 


OJ *43 


r- in 


in m 


CTi f) 


r^ o. 


cr r- 


a ir 


oc cr 



OJ i* 


■J} o>* 


ao m 


oo r^ 


C oc 


J cc 



< *e 


^* s* 


»- o 


m ^h 


< o 


■^ OJ 



en .-< 



r- o 



o *« 
in o 
oj r^ 



in «* 
in --> 
00 oj 



oo X 

-o tn 



a 

3 





13 




0L 


>- 


CO 


13 


^} 


O 




3 


z 


rc 


tri 



3 



X 

a 



nr 

Q 



lu 

2 

oa 
u 
z 



CO 

5" 



o 

z 



o 



or 

a 



4-36 



139008 IViOl dd lN334Hd 



<-3 



5 




LU 

ii. 


_1 




s: lu 


w 


<L 

< s 




< u_ 
cc 




1 1 








_l 














l7j 


n 




a. o 












« 




O LH 










M-i 


i/l 


i-i uj 




1- 


n 


> Li- 


O LO 






ce D 




t- 


> 


LU CC 


u_ cc 


l/l U. 


in a. 




aagwnN xaasii 
4-37 



TABLE 4A - 4C 
BIOMEDICAL AND BEHAVIORAL RESEARCH FUNDING IN 43 PRIVATE MEDICAL CENTERS* 

Averages per center in thousands of 1964 constant dollars; NIH R 5 D Deflator 

1968 1970 1972 1974 

Regular Research Grants 2317 1909 2217 2541 



Program Project and 622 652 1137 1412 

Center Grants 



Contracts 1 1 10 931 



Training Grants 1343 1250 1185 1246 

and Fellowships 



Total from NIH 6894 3895 4906 7030 



Total from NIMH 631 591 490 489 



"Data from NIH/IMPAC File 



4-39 



* 


oi 


If) <?« 


o *e 


on « 


r- x 


r^ 


o 


OPJ 


■rH ■* 


otu 


r^ xo 


i 


-■« 


in oo 


*-* o 


wff\ 


m -• 


f) 


•» 


• 


• 


• 


• 


r- 


»-t 


in 


o 


01 


■* 


(p 




* 


«H 




m 



0% W 


-i j* 


on ^ 


om 


OJOJ 


- < - * 



I 
•1 



On ** 


f*l cw 


-» *« 


m x 


r^ in 


r*>on 


0"! -O 


>o on 


-* 00 


■^ »-* 


^> 


ei ** 



O 1 >* 


o X 


oo r- 


OfU 


rvj --• 


t* 00 



UJ 

-I 
m 
< 



r- 


rt 


* -~i 


,-i<T> 


On oo 


-* o> 


m oo 





o 


-* o 


*H >0 


r*. 


m it 


c 


.-. 














r- 


■r~* 


~» 


.-• 


on 


00 


m 


(T-. 




* 


v-4 




n 




.-i 














^ 


-O 


en x 


* *< 


o *« 


OJ »< 


o X 


IN 


»* 


■H . < 


0*1 (VI 


on ^» 


-♦ m 


rvi in 


■ 


en 


* 00 


m 


oo 


m o 


fU 


< J 




• 


• 


• 


• 


■ 


r* 




in 


o 


(T 


* 


ru 



r^ x 


o X 


in i^- 


on rv 


fV (T 


a 



in *< 


r^ *< 


•rt CTl 


00 O 


oj on 


o 





-JO 


in 

a. 

UJ 





tv i< 


m *e 


on X 


*-• x 


o « 


on ■* 


on o 


r* r*. 


m o 


m ^ 


ro oo 


m 


r^ 


m m 


o 



o ** 

oo o 

moo 



m « 

On aO 



O0O0 

o 



oj on 



oj in 



o o 

OJ o 



• 


• 


Q 


n 


«-l 


OC 






rs 


OJ 


<D 


03 






ii 


^ 


OJ *< 


on ^e 


■< 


<i 


on m 


r^ so 


i— 


r 


^ r- 


M 


<* 


< 



n 




_l 


a 




< 


z 




u 


LU 




n 


rr 




in 


LU 


<r 


i 


H 


O 




^ 


ON 


LU 



o 



?< on « 


^r *< 


in ** 


ao x 


* on on 


r^on 


in * 


*-* rv 


\l o 


o 


OJ o 


OJ 



Oj M 


*M »« 


m #* 


* >* 


o^ at 


*c en 


on oj 


»^ r- 


^» »H 


--. rn 


(* a. 


if 


o. 


O. ff 


-* 



00 -^ 


r^ oo 


t- 


^-*oo 




«-« 


•4 


e 




o 


fl 


o 




m 




r- 




ft 


OE 


IT « 


o 


x 


<t 


sO 4 


r- 


K 


t- 


— o. 




r^ 


< 



o X 


0C *5 


f^ « 


ID *< 


o « 


-c o 


f^ 1^ 


r^r^ 


m tvj 


OJ o 


(»: *£ 


■-* 


o 


a lc 


<c 



ri mm 






< 
o 



0, 



>U X 


o *< 


t*> *< 


OJ o-f 


00 X 


in * 


oo on 


in ^ 


f*l o 


*-* t^ 


m oo 


r- 


*-♦ 


oj r^. 


1 



0~» i< 

aj oo 






3 



_1 
3 



z 


o 


o 


111 


3 


3 


1 1 


>- 


»- 


J 


CO 


U) 


(1) 


_1 


-J 




< 


< 


1 


u 


I > 


*J 


•-• 


— • 


K 


o 


o 


O 


iLl 


UJ 


►— 


r 


5" 



4-40 



SINSaniS "IV101 dO XN3D!d3d 



I 
< 




(_> 


















■a: 




cc 










l/> 




u_ 




o 

U_ 






-J 
< 




H 
Z 




ca 










V- 




LU 




•z=*" 




Q 


III 




□ 


ir 


e i 


_l 


■a: 


l/l 


< 


< ) 




H 


hi 


_) 


< 




er 


O 


7- 






o 


1- 


H 


1 


Ll> 


_J 


UJ 




i 


T 


U) 


O 


^" 


Y- 




►— < 






=> 




1- 


UJ 


■z. 


O 


u 




CO 




LU 




sr; 


UJ 


2 


l/> > 


1- 


_i 


H 


V 


i 


o 


n 




H 


z 


< 


< 


H 


i 




X 


o 


—I 


UJ 


i j 


_) 


_l 


o 


_J 


o 


>— < 


3 


Q 


i — i 


O 


=D 


Q£ ■* < 


< 


U) 


C) 


ID 


o 


< 


O 


Z ^ U 


UJ 


< 


< 


h- 


UJ 


Lt 


<. 



(J 
< 



ix 

LU 

m 



< 
> 






LU 



ca 
■a; 




aagwnN xscini 
4-41 



f^. --I 






oo oj 



(T\ in 



r^ in 
c-»r*- 



oo *< 

m in 

oj --* 



in *? 

in -o 



oj — « 
— • in 



o ** 

OJ OJ 

m in 



o *« 

-O OJ 



co *■* 
to *c 
oj in 



-* *< ox 

oo n mm 

oj no oon 

«-« m »-• r- 



<r x 


r- x 


i/i x 


oo x 


00 X 


-^ .* 


00 o 


o oj 


Ovo 


co -* 


01 r^ 


O 00 


«- o 


r^ en 


•H <i) 



oo ro 
--• in 


-* x 
oj oj 

J.) --o 


in x 

in ■-• 
oo r^ 


*© x 

■* in 

(VI *-i 


O X 


in x 

in 4 
in m 


o 


-< ^ 


m in 


CVJ o> 


m 


r4 *C 



in x 


O 


x 


•o X 


1^ 


X 


01 X 


r- oj 


r^ 


<i 


oo CVI 


III 


< ) 


oo en 


IT Ifi 


o- 


* 


* IT. 


J 


J 


■» oc 



*-• X 


CO X 


sO x 


no x 


--H CO 


m in 


in — < 


r- oc 


J- 0" 


OCT 1 


a m 


r- r^ 



ho rs cu 



--• .-• r^ 



o x 

•-• jj> 
r^ co 

m 



< x 

►- o 

< o 

o 



00 X 

oun 
oo in 



O X 


OJ X 


in x 


l*) X 


*-* X 


in x 


in x 


JJ 00 


n OJ 


O J» 


-i o 


co o"i 


oj a 


m 41 


co o 


in * 


oo in 


OJ OJ 


I-- w 


in -* 


in oj 



d) 



< X 


00 X 


ox 


O X 


<i x 


^o x 


m x 


o\ x 


OJ x 


Oj x 


1- o 


00 *-* 


*■ m 


00 -H 


oj in 


in oj 


oo r^ 


r- h 


oj in 


f^ 00 


< o 


en * 


r- ro 


H .* 


o -o 


OJ OJ 


r-. in 


r- oo 


r- oo 


in in 





h- 


<s 


CJ 


CI 


o 


O. n 


>-i 00 




a. 


en 


OJ 




CD 


1-1 






< 


r-t 






z 








U 
















O 


















oi 


OJ 


m x 


< X 


■JJ X 


cri x 




I 


jj 


LO 


OJ OJ 


t- o 


~o in 


00 nO 




o 


• 


as 


*o •** 


< o 


OJ oi 


OJ -* 


o 


a 


oo 












z 


< 


<c 


V 


m 


o 


OJ •-. 


--. 41 


'- , 


UI 


(J- 


■H 




CD 


•-* 





— -H oc 



o « 


in « 


OO 5< 


O 1 * « 


4 


J* 


00 5* 


o o 


00 CPi 


0\ "H 


OO O 


(A 


m 


oo no 


Ol 1ft 


in *o 


•O If) 


vO -* 


-* 


^»- 


o * 



< X 


O X 


.») X 


o X 


00 X 


o X 


* X 


en x 


>0 X 


oj 


X 


00 X 


»- o 


O 1^ 


-* oo 


CO 1^ 


o in 


O CO 


o r^ 


•-i 00 


o •-• 


r* 


'XI 


co en 


•< o 


■* OJ 


j> m 


o in 


en * 


m in 


CO *) 


jj ry 


*a oj 


CO 


O 


9\ P% 



^ H JJ 



o 



in 

or 

in 

H 



4 X 


-C X 


■O X 


CO JJ 


m r- 


00 01 


^ cv 


c c 


cv oc 



r- en ^ 



•f x 
ir i^ 

*- oc 



%C X 

in in 



— 1 * 


< *? 


in ^« 


xO *? 


oi « 


o rv 


t- o 


r- r^ 


r- o^ 


CO OJ 


s£ r^ 


< o 


-- •* 


-- m 


or a. 



r^- x h« 



^ X 


^ X 


JJ X 


in x 


CO X 


in xO 


oo ^f 


>0 41 


J- ^H 


r^ -o 


a oc 


oc 11 


a. cr 


ic r^ 


r~ oc 



111 

< 

a. 



» r^ 


X 


< 


X 


+ 


*C 


^r 


X 


00 


X 


<\l 


X 


-r 


X 


CO 


J# 


c > 


X 


CO 


X 


en 


X 


en 


<* 


rvi <i 


f\l 


i- 


o 


m 


J 


jj in 


in 


JJ 


m 


<1 


o 


in 


(VI 


J 


.-1 


cr, 


cu 


r- 


(VI 


^i 


rvi 


r* 


xi in 


OO 


< 


o 


en 


O 


.-< 


oo 


01 


J- 


m 


IS 


CO 


o 


0J 


Ol 


r^ 


r^ 


"-■ 


UI 


— 


JJ 


r^ 


01 


* 


m 


o 


-H 


PI 


,, 


Is 




^> 


^) 


1U 




OJ 


T-t 


oo 




* 


*^ 


r- 


* * 


r* 




-» 



J 



CO 

o 



< 



ui 






en 
O 

III 



< — • 



CO 

1 

CD 



4-42 



iBoang ivioj. do XNBOtQd 




s 


I 


> 

V 


N 1 


\ 


k^ 




/ \ \ 


X 


/ \ \ 


\ 


( Vi 


\._ 


V \ 


-I 1 ■» 1 


\ ) 



aaawnN xaaNi 
4-43 



TABLE 4A - 5C 



BIOMEDICAL AND BEHAVIORAL RESEARCH FUNDING AT THE 21 MEDICAL 
CENTERS RANKING IN THE FIRST QUARTILE OF RESEARCH INVOLVEMENT* 



Averages per center in thousands of 1964 constant dollars; 

NIH R&D Deflator 



1968 



1970 



1972 



1974 



Regular Research Grants 



4809 



3477 



4046 



4559 



Program Project and 
Center Grants 



1600 



1458 



2632 



3205 



Cont ract s 



20 



1694 



Training Grants 
and Fellowships 



3397 



2262 



2202 



2332 



Total from NIH 



8993 



7763 



9546 



12,416 



Total from NIMH 



1017 



1040 



826 



817 



*Data from NIH/IMPAC File 



4-45 



r\i x 


ro x 


*o X 


•*> ^ 


ox 


m so 


00 o 


•a m 


en o 


■JD OJ 


to r*- 


r« O 


«-i 00 


-* OJ 


ro 



O X 


o0 X 


nj ir> 


OJ if) 


■* r^. 


*-* Q0 



O OJ 
to •-« 



ro on 

OJ O0 



0> X CT> X 



on 



-i x 

oj •* 



oj in 



LU 

-I 

QJ 
< 



Oj If! --< 



^ « 


00 J* 


>o ** 


in in 


ID 00 


•« •"! 


-H ^» 


-* r- 


o 



(X) « 

-0 OJ 



O X 


*- x 


TH X 


ro x 


oj a* 


Ifl 0J 


JJ aG 


PI O 


T*r» 


-* OJ 


* m 


--4 O 


H Ifl 


-* o 


r^ 



OJ •* 



O 00 

■^ If) 



■* x 


00 .X 


m x 


^» X 


OJ o 


ro -• 


oj r- 


OJ *-* 


* if) 


*-* OJ 


^ OQ 


■* If) 



00 .-< 

o-i in 



□ 
or. 



or 

LU 



3 
O 

I 

u 

a 



OJ 5* 


-JD X 


rv x 


-^ ** 


on x 


r-. 


oo X 


o-* 


-* m 


oj on 


m oj 


ro on 


If) 


lo ro 


* n€ 


»<« oj 


^ ^ 


* -^ 


OJ 


m 


■OJ 01 



0J a* 

oi o 
on o 



o ** 

.-. oo 

<-* nO 

o 



-» 0> 

ro ro 



fu ao 






Q 
LU 



LU 



IL 

o 

10 

Ul 

o 



UJ 

> 

< 



p- »- 



mo -h -H 



0" •- 



r»i o-, 



on m 



If) X 


if) x 


OJ OJ 


m m 


rr o 


X 



in « 
ro no 



-r-l X 


ao x 


O tfl 


r*l oj 


m so 


OJ 



-C X 


V 


X 


■< 


<a 


oj r- 


cr 


OJ 


1- 


K 


a c 




< 


-a 


<i 


• 




m 


C 


r 


V- 




<X 






r- 




fV 


a 
7 


Z 


Lf) X 


4 


X 


< 


< 


o~\ on 


oo 


o 


l- 


1- 


-- r* 




o. 


-a 


** 


• 




• 


f 


Q 


on 




o 






•4) 




m 






o x 


v-i 


X 


•4 


< 


00 * 


00 


Lfl 


h- 


»- 






tr> — ■ 



lu or 

■Z. LU 

•- UJ 



4-46 



siN3anis nvioi do iN33a3d 





Ct Q 

LU ZD 



J < < 

- CD Li_ 



< < 



Q < 

Lu cr 




d39W0N X30NI 
4-47 



oj tt 


•o « 


OJ i< 


^-» « 


-* o 


m r- 


* o 


o * 


so ro 


oi as 


•o o 


<n * 



■^» •« 


*-i x 


ru x 


cn ;-« 


o x 


r*- x 


cn m 


orv 


r^ r^ 


^-t N© 


r- en 


oo r-~ 


%o r-- 


«-» *o 


r-* --• 


oj *-* 


<T\ -* 


r-- oj 



n£( w-* --» 



I 
< 

I 

UJ 

-I 

CD 



TO X 


-* ** 


r^ x 


*-* x 


r-- x 


m w 


m x 


OJ J* 


Cn oa 


-* en 


in r- 


oj m 


in oj 


<T\ »-« 


OJ sO 


oj en 


in *-* 


in oo 


o -a 


O 0J 


* OJ 


o o 


■^ 00 


oo r-v 



IT ->* 


<\ 


X 


•xt en 


en 


m 


oo o 


OJ 


■■•-• 



01 T^rH 



oo 


m « 


00 J< 


-» « 


00 irf 


CI 


* 00 


oc JJ 


■o o 


oo r-. 


%C 


in cn 


<P f\J 


-« <v 


cr in 



NO "rf 

O -H 

in r- 
C) IT) 



00 S* 

I"- o 



in in 
o r^ 



in o 



--• o 

(T o 






t- o 
< o 






-t « 

OJ 01 

in in 



-r-t M 


en * 


oo in 


oo en 


en »-« 


o 



*a ** 


00 i* 


o « 


.-. M 


— 01 


^ in 


oo v0 


in in 


O 00 


oo ro 


r-* o 


in o 



rvj m 


o 

z 


-i OJ 


0J 01 

r-t 


r^ 


oj m 

OJ 


o 


»H CO 


-~< r~ 


<o 


-B 


-* OJ 


"0 01 J* 


< X 


o -* 


■* d-f 


00 *f 


o « 


n « 


00 .•« 


o « 


LO X 


01 i* 


o ;* 


o cn r^ 


t- o 


-O vO 


■* -+ 


J> 0J 


<t <r\ 


•a •-• 


01 w 


00 00 


•-I OJ 


o r^ 


1-^ 0J 


yj 4» m 


< o 


* oc 


OJ OJ 


o in 


n i0 


.-» m 


01 — ■ 


o r- 


00 -O 


oc in 


oj m 



*-• ■** (\j ac 





D 




o 




I 




u 


o 


cc 


z: 


< 


i — i 


UJ 


o 


</) 


z 


UJ 


z> 


cc 


u. 






Q 


CE 


Z 


UJ 


(^J 


H 





N ^* S< 


< « 


•H 5* 


00 >? 


OJ « 


en ** 


■* •« 


00 « 


o « 


O i* 


r*> « 


tn x 


J- -r-i eu 


1- o 


-» O 


in m 


nO -h 


--• en 


in -* 


O-H 


en oo 


o oo 


en -o 


oj m 


* -* -sC 


< o 


in w 


DO OJ 


en ^> 


m <n 


■t-* TO 


O-oo 


r* o*i 


«o <n 


r^ en 


t4C0 



■rH --■ OC 



n ox 


< x 


CT> X 


•^) X 


^0 X 


rv x 


OJ ** 


0J X 


f*^ X 


^r x 


r^ x 


* x 


n o * 


1- o 


OJ TO 


0J O 


oo in 


*o o 


rv ^-i 


* 00 


m r-- 


00 TO 


in --» 


r-- ^ 


a -* rw 


< o 


sO OJ 


i) OJ 


r> to 


sO TO 


^ >o 


«0 OO 


*-* ■* 


|v TO 


ia oj 


en «-* 



< 


1 


III 


< > 


* 


H 


t-H 


« 


/ 


Q 


0i 


UJ 



O K 


■a iX 


— "■* 


c »« 


a 


m 


IT *« 


-- *« 


01 U 


»- o 


r^ ^ 


^H *C 


•A 


o 


r-« — H 


O --i 


P* •- 


< c 


<\ o- 


ur oc 


r> 


c 


tr oc 


o, «- 



o w 

OJ --• 

-t ir 






1 -o « 


< >? 


1^ « 


m -tf 


00 «f 


in •-<; 


o « 


O i* 


^ « 


O « 


in s< 


o « 


JO 00 -^ 


1- o 


oo m 


* i-n 


in in 


in <o 


r-t 01 


in * 


o at 


00 •-■ 


«H 00 


^ 00 


»- fT Ou 


•* o 


•» 0- 


^ cr 


r* cv 


— m 


CV C»t 


cr cr 


•- ir» 


un m 


i^ r-- 


r- a 



L9 

< 

a. 

UJ 

> 



X) * X 


< X 


rv »< 


r- *« 


^ *< 


en x 


r^ x 


en *« 


<J\ -K 


-• *« 


00 X 


TO it 


* r- m 


►- O 


en oo 


in r- 


cn to 


00 oo 


■C 00 


r-- o 


00 43 


^* in 


r- to 


* --» 


4" TO ■* 


< o 


ocn 


TO O 


in o 


cn m 


■r-* en 


ao -* 


O 


in o 


•o o 


v0 vO 



ID 

a 



z 



ii 

r 1 
) 

CII 

z 







in 




X 


l> 




u 


z 




tr 


3 




< 


U 




UJ 






C/l 


i 


l^ 


LJ 


u 


o 


or 


Or 



T 
U 

or: 






llr 



o 



u, 

2 



3 



<n 

z> 



a. 

on i- 



cn 

~> 

y 
in 



eg ■-• 



4-48 



laoana nvio± jo iNsoaad 





l;l 


1/j 


1 i 


in 






> 




ir 




l: 


■ r 




Q. 



CD 



ex: 



to -l 

SL < 

< 2 

CE O 

O (/J 

tx. to 

CL UJ 



< ct on: 

cc < uj to >-■ Q_ 

LU LU I— Q > tO 

o to < ~z. CT O LU 

LU LU I— => LU ^ LU 

U_ CC tO Ll tO < Li_ 








\y 


* 
* 


\ *\ 


& 


\j 




\t 




A 


« 
<» 

^ 


A 




\ ^ 


s 
s 


\ M 




1 H 


^ 

* 


1 !i 



dBGWON X3C1NI 

4-49 



TABLE 4A - 6C 

BIOMEDICAL AND BEHAVIORAL RESEARCH FUNDING AT 65 MEDICAL CENTERS 

RANKING IN THE SECOND, THIRD AND FOURTH QUARTILES OF RESEARCH INVOLVEMENT* 



Averages per center in thousands of 1964 constant dollars'; 

NIH R 5 D Deflator 



1968 



1970 



1972 



1974 



Regular Research Grants 1271 



1056 



1179 



1456 



Program Project Grants 294 
and Center Grants 



342 



528 



631 



Cont rac t s 



17 



679 



Training Grants 
and Fellowships 



508 



674 



658 



683 



Total from NIH 



4208 



2173 



2575 



3864 



Total from NIMH 



365 



317 



264 



275 



'Data from NIH/IMPAC File 



4-51 



crt m «h 



■ i - 1 



CO ** 

■r-*r^ 

o 






o io 



CT> *! 


ID Vf 


O i* 


--> ** 


4> If! 


4> O 


m ov 


r^ <r» 


* 0^ 


nj <n 


■rtOJ 


m -^h 



-ID o> 
OJ o 



w-t » Ol 



n. r- -* 



00 i< 


* ** 


oj w 


»-. in 


-* r- 


Oj nO 


~-> »-< 


*r in 


CT- 



o*^ ■«-# 



* *? 


<j\ »* 


.-* 


m 


^» r* 


oc o 


*t 


00 


o. o 


<r 




if 



*-i 


^-< 


* a* 


OJ *< 


OJ ** 


00 j* 


a\ « 


r* 


ro 


OJ (xj 


in (•". 


—< r- 


■* 41 


-H ■* 


i 


H 


* in 


oj ro 


^ m 


m r^ 


r- 


o 


•a 


• 


• 


» 


• 


• 


r- 


-^ 


r* 


OJ 


an 


o 


--H 



•jO »* 


4) i< 


■ u Oj 


00 — * 


OJ 0\ 


in 



r^ « 


o « 


00 « 


00 .-* 


m « 


o m 


00 * 


O Oi 


OJ *JD 


m oj 


j- *a 


—• en 


,-» ip 


m ct\ 


OJ 



-* t< 


OJ a* 


«-* m 


00 sO 


oj m 


-O 



3 
O 



OJ ** 


* 3* 


4) « 


4) x 


vO m 


■r-i m 


O 4> 


OJ 00 


^ o 


.-•oj 


pi »-* 


in 



Ol X 



□ 




111 


<r 




U) 


7T. 




111 


UJ 




cr 


a: 




o 


LU 


sf 


/ 


<r- 


N 


OJ 



0^ i* 


-4> *? 


* *« 


<-• J* 


l*» ^« 


4} -O 


a* oj 


00 *-< 


m in 


'* US 


H 


ai o 



CO O 



r* a> »- 






o 


i 


.-= 


|s 


CT 


O 1 J* 




^r 




vt 


00 


43 C3 


—1 


to 


(/l 


1 


€X 


o- %f 


< 


a> 


a 


c 




• 


u 


— 


LU 


X 




r- 


1—1 




H 


a 




* 


a 




Z 


«H 






UJ 




LU 









in w 

■rt 00 



m *« 


o « 


a\ « 


m *? 


r^ *< 


•jj in 


4» m 


ffl oo 


O") 00 


--« OJ 


m ur 


— -* 


41 


a o. 


c 



• 


o 


c 


c 


.- 


OC 






l^ 


a 


a. 


a. 






Z 


Z 


00 *< 


--* a 


< 


<< 


•* --* 


rv oo 


i- 


c~ 


~- If 


•* 


■a 


<t 



r-t ox 



o« 

OJ OO 



in x 
m oj 



2- 



Z3 



Z 

UJ 

o 



< 
_l u 
< — 
t- o 

tD UJ 
I- 3E 



a 



4-52 



siNaanis ivioi do iN33a3d 




.1 M.I.I I I . i . 
















(=1 



UJ ^ 

Q 1- 



Q < 

UJ cr 



LU 




usswrm xboni 
4-53 



in i< 


OJ X 


on m 


* m 


*o cn 


<T\ O 



r-*. x 


r*; x 


.-• x 


■r-» x 


m x 


sO X 


0"> X 


ro X 


*-( *-« 


-* *-• 


nO cvj 


in in 


^ (O 


aj j- 


r^rr) 


so r- 


IT) * 


CT» sO 


in 


•* CO 


in ■— 


>D ro 


CO OJ 


cn «o 



tr ** s* 


m x 


in x 


oj x 


fV.' X 


rv ^ 


o X 


v£ X 


IT X 


a x 


r^ x 


cr x 


n inoj 


Cn ^J 


■* in 


OJ c*> 


-n m 


oo in 


sQ ^-t 


sO O 


U . • 


cn ro 


-* oj 


m r^ 


■a o no 


00 nO 


in oj 


in o 


* in 


--* r-. 


in 


-* o 


cn r- 


«C Cn 


0J T-* 


-r-. r-v 



.* «-H T-t 



OJ X 


m x 


in x 


o x 


— • x 


nj i* 


^ x 


^ X 


rn X 


OJ x 


en x 


r- o 


oo rvj 


on -* 


m r- 


■* in 


in ru 


oj in 


on m 


J- -rt 


in *-• 


in r^ 


if, ^ 


n ^c 


cn *-* 


in oc 


<. a 


CT> 4 


-- 1H 


-* •* 


* <A 


IT. Oa 


in ou 



< 


M 


00 5* 


1^ * 


i^ x 


in x 


i- 


U 


cn in 


m •£ 


on -a 


ao on 


<t 


o 


«-H -* 


* in 


-* in 


r^ cn 



in x 
orv 

r-. ■* 



in x 

en -4 
cn ao 

--. OJ 



OJ x 


* 


X 


in ao 


41 


»-t 


OJ sO 


OJ 


O) 



OJ O *H ^ 



tu 


o 


(VI 


r*> x 


<* 


X 


cn x 


an x 


on x 


^ X 


nO X 


in x 


r*) x 


O X 


* X 


in x 


_l 


r- 


OJ 


oj en 


i- 


o 


^-i 00 


<H »-t 


wt cn 


r^ oj 


^r o 


m ao 


on in 


r- o 


O -H 


OJ --I 


«-< 


• 


OJ 


* in 


*t 


o 


o on 


Cn OO 


in ■* 


oo m 


in 


*o oo 


o oo 


* --i 


00 I-- 


o -< 


H 


cr 






























or 


sC 


cn 


^ 




a 


— a 


— i o 


in 


•h a 


o 


^ 


n »— 


in 


OC 


*-< ■-! 


< 


en 






o. 




r-< 


OJ 




OJ 






•-t 






r-t 


O 


«-■ 






^ 

























IB 


< 


z 


HI 


1— < 


(/) 


Q 


UJ 


Z 


a. 


=> 




li- 


a 




tr 


CC 


(Ti 


UJ 





< 
o 



< 
CJ 
< 



O X 

o ^ 



OJ X 
rco n0 



^r x 



sO x 

m ^ 

a in 

OJ 



in x 
m T-* 



-t x 

oo m 
■■o in 



r-i rv X 


-4 


M 


>0 X 


rv x 


•o X 


O X 


■^ X 


OJ X 


* X 


nj 


X 


ao x 


m 


x 


c --« CO 


i- 


O 


m «^) 


^) r^ 


O m 


00 O 


so n 


oon 


O OJ 


oc 


•JJ 


OJ 00 


cn 


m 


n ^* <n 


-* 


CJ 


*-• IS 


a\ > 


in o 


n ^* 


r^ 


iO --< 


^» 00 


* 


r*i 


GO 00 


r. 


LB 


o ^» 


o 


•-< m 


^ m 


X) 


OJ 00 


o 


r^ 


■* 




m 


en 




m 








"~ 


Ov 




a 


















r- ox 


<x 


M 


OC X 


<. 9* 


t- X 


4. » 


r* x 


c X 


4 X 


* 


X 


* X 


r^ 


X 


O OJ OJ 


t- 


( » 


o in 


*-« f*) 


m >o 


o on 


m o 


r^ m 


<c o 


rv 


*> 


I s - o 


ra 


•A 


a ^ — 


<i 


c 


cr *- 


OC «- 


•* O- 


o* C 


r- 


^ h- 


- c 


f 


u 


*- O" 


r* 


at 


DC If 




CJ 


a 


-- a 


u 


a. oc 


c 


if 


o 




-4 


^- ^ 




of 



<o 


r. 


OJ *? 


ji 


tn 


O -0 


• 


c 


4 -t 


id 


q 


• 


-u 


Pv 


in 



r-v x 


on x 


oc X 


OJ x 


oj in 


i*- r-v 


Ol f» 


o o 


0C * 


r- a 


^* »- 


c a 






rf rf rv o 



00 *« 


IO 


w 


-» m 


11 


in 


^ IT 


IX 


: T 



< 
o: 

uj 



o cn 
^ oo 



« o 
o • 



r^ x 


w* X 


OJ o 


oo m 


vO 00 


-* o 



h ^ m 



O X 


ox 


o X 


cn x 


r-x 


oo 


X 


r^ oj 


in ao 


0O CO 


OJ OJ 


m co 


in 


<X) 


o 


* in 


*-t 


m oo 


rvr^ 


r^ 


o 



(1 



UJ 

(j 

> 

m 



CO — " 



I 



r 
u 



n 
(3 

a. 



z 



CO 

3 



*-• 0C 



CO —• 



4-54 



j.3sana nvioi do iNBDaad 











1— ( 


co 


O 


IT O 




CD 


h- 


Z. 


o 


CD i-i 






< 


O 


_J 


□ CO 




O 


q: 


D_ 




or co 




z. -z. 


LU 


CO 


Q 


D_ LU 




i-f o 


Q_ 


X 


2 


Lu 




t- -« 


o 


—1 o 


< 


lu o 




< -I 


h- 


< cr 




u cc 




CE _l 


_l UJ 


a: < 


LU I/) 


i-i a. 




LU >-• 


< o 


LJ U 


1- Q 


> 


LO 


d_ s: 


1 ■ Q 


Q 1/) 


< 2 


cc q 


hi 


a 


O =D 


LU UJ 


h- => 


lu z: 


III 


\D 


h- CD 


Li. Ol 


CO L_ 


co < 


Ll 




H 


L_ 


</) 


D_ 




X w 




aSBWON X3QNI 



4-55 



m *■? 


* *e 


CT\ ** 


oj *< 


no ** 


in in 


OJ ^0 


o * 


CT\ H 


--I ■* 


m <<l 


OJ <J\ 


--•nj 


oj r*. 


in 



•r* -* *e 



o) « 


w-t 


•o 


oo in 


■v 


CJ\ 


r*. 




in 



< 



o J* 


r*l *• 


in a* 


O irf 


r-v w 


(*1 OJ 


«-t a 


<T\ oo 


oo rv 


w-t 0J 


in co 


Oj 00 


m 


OJ r^ 


in 



0O <C 



oo m 
o 



ID 
-I 
CD 
< 



*• a* 


o « 


in ** 


o-i *o 


v0 ^ 


th r*> 


«-« 


Oo m 


-* 



in ^* 


o ** 


r^ o> 


oc <r> 


-- in 


■* 



SO i* 


H *< 


-o r^ 


x ~o 


» -i 


•-• *0 



oj k 

-i 00 



oj «-!»-* 



o *< 

oo -» 



r» ^f. 


or* s< 


r*- a* 


OJ X 


<T\ -* 


oo in 


*o r- 


OJ ^o 


-•in 


00 


OJ -* 


OJ 



•O it 


OJ *t 


tn c\j 


rv i^ 


^■m 


«/D 



o 



or 



OJ « 

aim 



CFl S* 
OJ oo 



■^ *e 


o .* 


0> J< 


OJ ** 


rv x 


«H PO 


r» oj 


r- <t\ 


h CO 


rt 


* r- 


•-• m 


a\ 


OJ --t 


flTU 



r* «* 


<T\ 


« 


< 


-* in 


J") 


«> 


1- 


--•m 




m 


< 


• 




V 


n 


a 




m 




i^ 




m 




in ^ 


* 


.w 


■4 


m oo 


*£> 


•^ 


*- 



<r 


O' 


L0 


c > 


.-• 


ir 


i— i 




UJ 






«- — M 



-- ^ oc 



UJ 




Q 


LL 


< 


O 


O 




< 


i/> 



CO }* 


sO ■**" 


OJ ** 


OJ *? 


0*1 i* 


0\ -• 


■* CO 


r- *-i 


on in 


AJ 


o- |n 


-- ac 


ac 


-- -* 


"■* 



<T> « 


i^ i* 


o X 


>o *< 


•o « 


Q0 -* 


m in 


NO f^ 


r^ ro 


Ol 


m in 


^■00 


00 


--* CT» 


f^ 



OJ *« 


.<) 


** 


< 


< 


Oh 


in 


rvj 


ft- 


1- 


•1 ir. 




4 


< 


•a 


• 




a 


O 


( 1 


* 




in 






•o 




m 


0J 




in ** 


rf 


M 


< 


< 


uTt o 


in 


(Tl 


►- 


►- 



UJ 

> 



o 
.z 

I.I 

>J 



I" 


1 


J 


a-> 


1 


ir 


Jl 


? 


<r 


iii 


^ 




iii 


K 




? 


1- 


i.i 


y' 


CI 


UJ 


_} 



J 

3 



Z 
( .l 
O 



-I u 

< « 

t- a 

CD UJ 



O 

3 



Q 



t- 


■-• 


c_J 


c 


I J 


z 


J 


in 


l.l 


1 




■— f 


) 


_l 


u 


lA 


< 

LJ 


in 


1 


•-• 


u 


t 


z 


-- 


I- 


■ — 


c/) 


03 


-J 


*l 



4-56 



siNaanis nvioj. jo iNsoaad 







01 o 

LU ID 

a H 






J »-i rD 


Q 


t-i O 


~H 


<(/)U 


ID 


o < 


C 


.a < < 


H 


lu a: 


O 


— CD IX. 


l/) 


2: 


LL 


— 


I/) 


2: 


LL 




1 


■ 

I 


■ 

1 
1 




u 


■ 







d39WHN X30NI 

4-57 



o h* s< 


m« 


o »< 


m w 


^« 


^ 


* i* 


n i* 


QO ** 


LO £4 


n *■? 


* *« 


■* *? 


o «oo 


u\ --• 


m o 


vO m 


* 


r\i 


m oj 


>*o oo 


o ■* 


OJ If) 


in <r^ 


ru -* 


r-. r-. 


~n in *-• 


-* n 


(J\ o 


■X) M 


"* 


LI) 


Or* 


>o 


•a in 


r*- m 


m r^ 


Lf) -*0 


if) *-* 



OC X 

orn 

in in 


•« X 

OO OO 


-o~-l 
o -o 


oj x 
in m 


in 


rv 


r-t rt 


OJ |v 

OJ 


00 X 


< x 
•- o 
•« o 


O X 

o o 
o o 


|v X 

*o Ol 

-. en 



CO r* 
O00 



flC X 


* >* 


,0 w 


1^ OJ 


m o 


co -i- 



o «-•.-< 



f^ ** 


if) *e 


OJ »* 


OO *? 


-O i* 


rv *«? 


* *^ 


*-< o 


-XJ 00 


o m 


o <n 


in r- 


oi if) 


■w- r* 


<r o 


r» 


e»i ^c 


«-• rvj 


<v o 


cr o> 


If) -H 



-< o 



[\J OJ X 


< X 


|v X 


-i X 


-O X 


m x 


00 X 


O X 


o X 


00 X 


rv x 


^ X 


a «-< in 


»- o 


-» OO 


O OJ 


X OO 


m * 


•-Itf) 


on ao 


* on 


OO Ol 


j> «-* 


on on 


-< * o 


< o 


O0O1 


J*- -* 


*-" OJ 


o -o 


-H -* 


on rv 


«o o 


oj in 


on >o 


on rv 



UJ 


o 


OJ 


*-* 


X 


< 


X 


in 


X 


rv 


X 


a 


X 


o X 


•-I 


X 


in 


X 


OJ 


X 


*H 


X 


e> x 


•O 


X 




IV 


in 


OG 


J- 


► - 


o 


on 


wt 


no 


r-* 


Jr 


in 


in 


in 


* r^ 


» 


in 


rv 


n 


on -o 


in 


m 


in 


CJ 




9 

01 


in 


on o 


< 


o 


*o OJ 


On 


m 


OJ 


•- 1 


oi in 


~* 


(X) 


in 


OJ 


4 


in 


OJ 


ao 


on 


^> 


OJ 


* 


a. 


rv 




ir 




r 




o- 


--• 


■a 




IT) 


m-t 


■r 




., 




l> 


._, 


ir 




r" 




o 




1*1 


< 


(Ti 








Ol 










OJ 








,H 












rl 








r * 






a 


.-• 








Z 












































X 


on 


on 


OJ 


X 


< 


X 


in 


x 


O X 


r^ 


^ 


in 


X 


OJ 


X 


-t 


X 


O X 


-i> 


X 


.., 


X 


* 


X 




•o 


in 


r« 


>n 


1- o 


aorv 


sO 


nj 


no 


in 


on 


-o 


co r^ 


j 


r> 


m 


•Jl 


o 


ai 


in 


J 


on 


J- 




■ 


TO 


oi 


f 1 


•« o 


rv ■*> 


iv 


m 


-^ 


in 




■«! 


* 


in 


j» 


-o 


ir 


* 


in 


»-< 


41 


^ 


■• 


< 


00 


















































• 



< 
< 



r 


00 


ao 


in 


w 


<* 


X 


an x 


on 


X 


en 


X 


on 


X 


in x 


on 


X 


rv x 


oi 


.t 


00 X 


on x 




•o 


O 


sO 00 


H 


o 


OU 00 


ao on 


m 


H 


on 


* 


•-•t^ 


oo 


si) 


OJ |v 


--i 


m 


m oj 


|v on 




1 

vO 


OJ 


on«o 


•s 


u 


r^ ~e 


10 


* 


,^ 


-< 


in r^ 


OJ * 


01 


l\J 


on oj 


O) 


•-» 


1^ OJ 


*-i OO 


z 


^0 




in 




o 


r, 


„ 


in 




m 


^ 


ill 


m 




-D 


in 




in 


ni 


OJ 


m 


o- 








a 




*r- 




a 








IV 














«- 














J 


































UJ 


r* 


■i 


o 


X 


<i 


J*? 


■C X 


-f 


X 


* 


^* 


o 


H 


IV X 


Ji- 


X 


>0 X 


f 


X 


-c X 


Ol X 




* 


P-v 


4 


III 


Y- 


<) 


r^ ^* 


-* 


OJ 


X 


ji 


on 


r> 


in -■ 


ll i 


111 


OJ >o 


m 


» 


in -c 


OJ o 


z 


• 


ir 


f 


Jl 


i 


r- 


ir c 


jj 


ir 


l» 


it 


if 


c 


■^ OC 


J 


«- 


-- o. 


rv 


o 


t-- ir 


•- c\ 


UJ 


<c 










































u 


•A 
O 


if 




■A 


a 


c 


c 


*" 


if 




^ 


— 


at 
o 


o 




01 


a 




* 


or 


a 


< 


"^ 








z 


































*-■ 


vt) 


ru 


in 


X 


< 


X 


m x 


in 


X 


41 


X 


m 


X 


00 X 


m 


X 


Pv x 


n 


X 


OO X 


m x 




•0 


in 


in 


O! 


1- o 


v0 OJ 


OJ OJ 


>f 


m 


^ 


--t 


in in 


oo 


in 


r* in 


on r^ 


Oj ^O 


*-* OJ 




a 


<* 


o* 


sfl 


<* 


(-, 


ir ir 


«-i 


n 


f 


nr 


-£ 


«. 


w- cr 


(T 


rl 


j» 


o. 


Jl 


If OC 


•- * < 


s. 


in 











































IS 
< 

UJ 

> 



O ~H w-l 



rv x 
iv o 

m so 



< x 

I- o 

< o 

o 

ID 

z 



~-i X 

O Ol 

in o 



rv x 

O 00 

on j» 



J- i< 


in •** 


(D « 


«£> --e 


CO O 


r^ m 


o m 


no m 


in <y\ 


~* m 


00 CO 


m 



JD X 



rv v-i r* en 



m x 

o ao 
•-■ o 





a 




ri 


1- 


OO 


1'J 


U5 


o 




3 


z 


CO 


0i 






15 


1- 



or 




a 


UJ 


en 


z 


> 


l- 


3 


CO 


U- 


U- 


o 


»— • 






ca 


f 


_l 




u 


** 


Q 


or 



CO 

X" 



o 

01 
Hi 



CD 

z 



3 



Q 

Z 



5 

(i 



Q 

Z 



4-5c 



lasang nvioi do xNB^a^d 



UD 

a 




LU liJ 

Li. cr 







d39<nN X30NI 

4-59 



aj 


r^ « 


r* x 


h- * 


m x 


m i* 


[U 


C\J X 


o « 


(V) 


ro m 


r* m 


r* ^^ 


* co 


OJ 


* 


vO OJ 


co r* 


■"• 


m co 


«-ioo 


oi 


m in 


in 


OJ 


.-i co 


"2 



If) i* 


IP * 


-* >P 


r- j* 


O 4 


If) f^ 


■-U nO 


«-« fT\ 


IT) 00 


.^,0 


o 


ro as 



OJ <r-t -r-l 



a 



LD 



* ** 


NO .-* 


--» 4* 


in ** 


*£ *< 


%o m 


oo rv 


no ai 


co 0> 


oj 


■* 0. 


-- cr 


-* 


Oj O 


r- 



CO 


*• 


--» x 


» 


* 


i- in 


•H 


sC 


m 



<o x 


i-- *» 


in « 


in « 


c> o 


»< o 


-o a\ 


•^ CO 


* CO 


r-i r-t 


CO 


<\j r- 



-o « 


CO X 


* 


m m 


-C 4] 


r~ 


-4 in 


* 





^ 


co ** 


0) ** 


m a* 


n ;* 


an 


OJ CO 


If) * 


00 >o 


^o 


CJ 


-Htfl 


o 


OU xO 


sO 



OJ -r-t 






3 
a 



%© i* 


ru x 


»-* *? 


* *e 


^} « 


r- * 


--•if> 


^) ,-• 


0> OJ 


If) 


oo m 


^-t ai 


OJ 


^<i o 


CO 



0> X < 

in a\ i- 



n 




HI 


ir 




Ifl 


Z 




LU 


LU 




II 


cc 




T 


LU 


<r 


H 


H 


N 


<r 


^ 


a 





oo ** 


CO <** 


ao ** 


O i* 


ao r^ 


or- 


* OJ 


ao if) 


oo oo 


*-*ao 


-■o 


•H co 



G * 

m OC 



If) *H 



in 


^ « 


m 


m m 


.-< 


•-> i» 



o 

or 



If) If) 

o- 



r» o w 



« in se 


If) S* 


If) i* 


-H « 


r\i on 00 


■* oo 


rv j- 


o 


[T dC 


if 


■- j 


(V 



ru *? 

oo *o 

if 







i> 




(i 


m 


o 


in 


<j 


o 


D 


m 




in 


*-< 




(\j 




< 




** 














< 


in 

LU 
















in 


r^ 


f^ *e 


*-• *< 


OJ « 


^r x 


OJ X 






4) 


in 


•C <Ti 


CO OJ 


* a\ 


sO xO 


m 




a. 


i 


■o 


m oo 


^» 


m 


^< ai 


m 



m ** 

» in 

in 



ui 



1 



J 


0) 


I 


>l 


(II 


s 


ir 


i, 


Z 





z 


U) 


UI 


Id 


o 


or 



3 

u 



3 3 






a' 



u 

< 

tn 



4-60 



siNsanis ivioi do iN33a3d 







<r o 




a 


o < 


_J 


< 


IL Ll_ 


< 


or 




H 


O 


Q LU 


a 


cc 


< O 


h- 


LU 


a e 




Q 


_I LU 


1 


2 






z> 


O O 


<s> 


1/) 


^ CO 


\- 


-1 h- 


!—• 


^ 


< -z: 


I u 


LU 


CJ LU 


CJ •-< 


Q 


t-« O 


< LO 


=> 


O =) 



C3 



a 

■SC 




d39W0N X30NI 

4-61 



1 



UJ 

-I 

m 
< 

1- 



o * x 


00 x 

OvO 
CO Ol 


o x 

00 O 

sO ^ 


in x 
o * 
in o 


ro x 

in ro 


o\ X 

ro rs 

OCJ 


oo x 

oo ro 
in 


r^ x 

o\ in 


--» x 

CO OJ 

NO --• 


in x 

in -* 
oj in 


ru x 

en o 

in ro 


tn x 
m ro 




■^r^ 


--* o 


*-• a\ 


m 


OJ 


o 


in 


CM 


— tr*» 


ro 


in 


m vo x 
* so ro 
n oo c\J 


o x 
in m 
r- in 


■* in 

CO ** 


in x 


s0 x 

o *o 
r^ ro 


rn x 
*\j ro 
<\jin 


-* 


in x 
>o ro 

o^ oo 


r*. x 

o o 

in oj 


00 x 

in •-• 
ro cvj 


in ■* 


oo ro 


sO in 


* 


w ,-. 


tH 00 


4- 


-* in 


o 


in 


m «-» 


•H 00 


n 


-* 



O it 


O J* 


-O it 


00 X 


4 it 


OJ X 


<n ro 


ro oo 


r* oj 


iv m 


ro in 


■* •-* 


m ■* 


r~ r- 


in oo 


-i a 


O -> 


(v> in 



ru j* 


* « 


.-i ro 


in in 


n on 


0"* ai 



in 



»- o 
< o 



<o it 
o oi 
•o in 



>o ro 
in ro 



ao ro 



ai en 
in -o 



^ in 



oo it 

in r* 

in oj 



^* -* oj r* 



oj « 

ro o 
oj 



O X 


ao x 


* rn 


in :n 


ro oo 


ao -4) 



O 

Z- 



UJ 

1- 



~-* 


J* 


< J* 


ao X 


o »< 


111 


ffi 


►- o 


«-t 00 


CT1 O 


Al 


-» 


t o 


in * 


OJ 01 



in m 


< X 


-O X 


-o it 


m O) 


1- o 


—• 00 


rv ^ 


tn o 


<t o 


OJ OJ 


o -« 



OJ X 
«H 00 

o-o 


00 X 

oj r^ 
oj in 


0"\ w 

in in 


in x 
-* in 
*o ro 


so in 
o ro 


ro x 
in * 
ru >o 


so in 


4 1^ 

OJ 


■^ 


"2 


a oo 


■r- r- 


TH OC 


^» 


O M 


ro it 
» in 

OJ o 


oj ro 


ro « 
ro r- 




tn --* 

4M sO 


4" X 

ro ro 
mco 



ao 



no •-. -i 



in 
nj 



«-* 00 
>o an 



l- O 
■« o 

a 



O X 

r^ oi 

-8 «% 



ao oo 
•1 .o 



in x 


•jj x 


in x 


•-> it 


no j* 


r- o 


in --i 


*o 1^ 


OJ » 


ro *-« 


4 us 


ro on 


«> ao 


r^ o 


<n o 



I 

-? 



O n n 



a\ oj 
ro oj 



< 



in 
< 

UJ 

I 
i- 
a 
a 
z 



f o. 



in ao 
o- < 



o >o 
o" tr 



o o 



j» a* 

m ^i 



00 00 



-~ O. 



»— 1 


OD 


•a 


•& 


ro 


<* 


-a 


„-* 


Ol i* 


■« X 


in 


«e 


*> it 


r- it 


at) 


i* 


O it 


in it 


IV « 


rv x 


^ 


<\J 


<c 


•it 


(71 


r^ 


► 


(i 


OJ 4T 


r- oj 


or 


m 


in in 


in oj 


(11 


00 


* r^ 


■O n 


OJ t4 


m in 


UJ 




■ 


* 


u 


■i 


«* 


o 


o. r- 


o- r» i 


ir 


ir 


o r- 


Ol 4 


a 


11 


1^ o 


l> O) 


c a. 


^••rv 


Q 


IL 


in 


« 




a 


n 


• 


^ • 


■ 




o 


^ « 


• 




• 


• 


■ 


^ • 


• 


< 





•0 


o 




III 




o 


»H ri 


o\ 




IO 


* 00 


ro 




ao 


i> 


r^ 


-■Ol 


-* 


O 




Ol 


--i 






en 




r* 








m 
















< 


10 

UJ 


" 








z 






























< 


in 


^* 


* 


« 


<t 


X 


ru »e 


* »« 


r> 


x 


in it 


OJ M 


in 


it 


OJ it 


oj *e 


a> « 


ao it 




X 


~o 


o 


w 


vf) 


t- 


r> 


« in 


in ro 


^4 


CVI 


ro ro 


00 O 


in 


-il 


OJ -» 


-o « 


0> |v 


O iO 




UJ 


s 


00 


in 


Ol 


< 


o 


•-• 00 


01 1^ 


■4) 


ru 


00 *-l 


ro a\ 


ao 


r-~ 


>o ro 


iO r^ 


ao «h 


^t «-• 




> 


* 


^ 




■ 


ri 


• 


^ • 


• 




« 


ea • 


• 




• 


• 


• 


• 


• 




< 


as 


a\ 




111 


aj 


o 




o> 




*o 


ror/i 
ro 


ro 




OO 


4) 


<0 


01 


J- 



J 
u 



_> 



l.l 

in 

a 
on 



(3 .-. 





(D 


L 


_l 




U 


< 


LI 


'r 


( J 


Z 


< 


<tl 


< 


UJ 


J 




in 




in 


U) 


CI 


1- 


a 


^ 


Z 





a 

3 



£1 



2 


<t 


u 


or 


r> 


Ul 


3 


o 



o — ■ 



4-62 



laoong ivioi do iN33U3d 





(J3 



i 

ca 



l±j 
<=> 






or 


n 




o 


,—. 




i i 


(/) 




01 


[/) 




u_ 


LU 
U. 




III 


r i 




f ) 


a. 






n 




> 




CO 


cc 


i i 


iij 


LU 


?*■ 


in 


if) < 


u_ 




aaawnN xsoni 
4-63 



o *e 


■* s* 


ro *« 


<j\ i* 


-*■ ** 


m o 


en -r^ 


«-* n 


OJ 00 


oj m 


in T-* 


»-»OJ 


*-* ■* 


* NO 


o 



r^ « 


o x 


r^ 


r\l 


Ol/l 


OK) 


oo 


OJ 


ro r- 


*-t CO 




JJ 



u 

1 



o X 


or x 


rn x 


r* x 


f» x 


00 


* X 


m x 


--< o 


ajoo 


Oh 


o-o 


m oj 


o 


OoO 


O •-• 


If) -J 


*-* 00 


^th 


-» <a 


OJ 


* 


ro >o 


^» m 



ID 



f\l 


rt 


0% ** 


sO *< 


vO X 


OJ ^ 


r^ 


C 


r- o> 


m r^ 


<r to 


a. o 


• 


•H 


* -* 


-~cn 


r- 


m ^j 



r^ 


--• x 


o X 


o 


|v 


00 00 


0^ it 


•c 


m 


OJ -* 


in 


n 





--> 


r* 


m 


M 


-r-t « 


00 « 


-H « 


OJ « 


a 


*o d< 


a 


<* 


o 




r>. 


r 


in 


u0 


OJ o 


00 -0 


on f») 


OJ If) 


r t 


OJ «H 


a\ 


90 


in 




■ 


m 


* 


00 


*-* 


o 


m en 


OJ 


en 


OJ IO 




■* 


r-| 




o 


















• 




• 






r- 






■Xi 


OJ 


0". 


m 


OJ 




*H 




oC 






a, 






* 


^< 




m 






r. 




OJ 






.-< 




























o 


m 


n 


„■*■ 


I s *. *« 


OJ ^? 


NO ^? 


r^ k 


n 


oj *e 


.-* 


.*■ 


< 




r- 


00 


m 


rl 


oj -* 


r- n 


* oj 


m m 


00 


O -* 


oa 


in 


H- 




■ 


<T> 


-* 


o 


*-icr\ 


m 


m oj 


r^ 


OJ 


Oj Ol 




-ii 


< 




a> 


















■ 




• 


O 




<£ 






4> 


<* 


r 


IT 


m 




«H 




<X 






O 






J 


H 




m 






r~ 




OJ 


oa 




*- 
























. ? 


UJ 





























a 
or 



*-l « 


OJ d* 


r- « 


*o S< 


^r ** 


«-* CT\ 


oo 


r^ oo 


»o o 


OJ * 


■* o 


** IO 


*G 


OJW0 


r*. 



nO ^h n 



o « 


<-i 


■»* 


< 


< 


o *-• 


0(1 


aa 


*— 


»- 


OJ OJ 




r* 


< 


< 


• 




• 


o 


n 


*-■ 




an 






r- 




OJ 


(D 


2 



O *f 


m <* 


sO ** 


* ** 


in j-e 


*H •-* 


OO 


•« ■* 


>« w 


■^ OJ 


* 1^ 


-•o 


rv 


oj r-« 


Q0 



Ol .* 

oj ^ 



< 


a> 


K 


( ) 


•-I 


LT 


>— 1 




U 



O X 


OC J* 


J» X 


O OJ 


en oj 


•C (*) 


^ ir 


O" 1 


OC 



ac IT K 

o. -- h- 



o *? 


in i* 


OJ s? 


^-» i< 


*-* ** 


u 


-*■ if 


CT> 1^ 


r- oo 


r^ ^r 


in in 


oj m 


rt 


4 (J\ 


c; rr 


o. 


ac 


a oc 


vC 


a 


«H 4 



< 



00 £« 


CO X 


o X 


OJ M 


CO « 


Oi 


CO X 


00 CO 


rv o 


in «h 


OJ ID 


*-i OJ 


o 


r*i oj 


m o 


in 


<a 


OJOJ 


in 


OJ 


--> * 



»-• o 



vo h ►- i- 

co < < 

• o o 

CO (D iD 

z z 



UJ 

51 



>- 
>- I- 



_1 3 
=) L) 



O 

z 



I ' 
fl 



III 


u 


It 


_l 


_i 


■> 


z 




3 


_> 


■— ■ 


UJ 


UJ 


u 


i ) 


)- 


•—I 


«J 


1 


< 


■ 


1 1 


z 


u 


u 


1 


to 


UJ 






1 




•— 1 


UJ 


ir 


-> 


_l 


L) 


5-: 


,•1 


u 


< 


en 


— 


UJ 



— . LI 



■-• CO 



4-64 



siNaanis ivioi do iN3Dd3d 




a: o 

LU => 



J < < 
- CD U_ 



t-H Q 

a < 







uaawnN xboni 
4-65 



I 
< 



CD 
< 



* 


o» 


OJ 


X 


* 


*< 


IN. 


x 


m 


X 


.-» 


X 


m 


X 


-* 


** 


a 


X 


ox 


.--« 


X 


sO 


X 


m 


X 


r-N 


«-l 


O OJ 


m 


m 


O) 


m 


r^ 


v0 


in 


fh 


m 


IN. 


^) 


:») 


OJ co 


^) in 


m 


O 


or- 


OJ 


J J 




in 


•-I 


-* 


m 


o 


cu 


en 


o 


00 


r- 


co 


00 


co 


OJ 




If) 


in 


VJ 


j- 


f > 


o 


r- 


in 


r-4 


0> 


a0 


CM 




m 




m 


--H 


--« 


cu 


ro 




OJ 


OJ 


en 




O 




* 


«-« 


cu 




in 




* 


r -, 


in 


m 


»-H 












«-i 




(V, 








»-l 












w 












«H 


-^ 




















































rr 


r 


,._, 


** 


tr 


*# 


or: 


x 


ir 


** 


co 


**■ 


on 


X 


a 1 


K 


rv 


X 


o> 


X 


n 


»* 


rr 


X 


n 


X 


r 


.-* 


m 


* 


* oo 


in 


en 


00 


i) 


-t 


r^ 


'X) 


OQ 


in in 


oo 


fVj 


oj in 


<-4 


•*> 


m 


m 


CO 


OJ 


I 

OJ 
|N 


IN 


<-n 


m 


in 


*o 


CO 


in 


in 


O 


CU 


o 


CO 


ro 




-* 


J- 


V* 


in o 


|N 


o 


m 


ai 


*o 


* 


,H 




(*l 




* 


_, 


,. 


m 


a 




Aj 


a 


o 




o 




4 


^ 


rr 




vC 




rvi 


H 


* 


(T\ 


»-l 












■r < 




0J 








OJ 












- 1 












-' 


""* 




















































OJ 


o 


m 


** 


to 


X 


01 


X 


in 


X 


t 


X 


Ol 


x 


■h 


X 


OJ 


V* 


CT> 


X 


4) 


--»■ 


m 


X 


If 1 


ft* 


r^ 


CI 


« 


O 


r- 


* 


(O 


Ol 


r^ 


r^ 


«-« 


vjfi 


sf) 


or 


* 


U) 


U) 


u 


00 


OJ 


(\i 


0" 


an 


n 


r\, 


o 


■ 


c 


n 


m 


r* 


o 


m 


r~ 


■H 


|n 


* 


r- 


J/ 


r* 




w 


01 


(V 


m 


•i 


43 


tC 


in 


o; 


rx 


in 


*-* 





















































fo —•cu 



^ x 
in o» 

CD ^* 



in x 


O X 


m x 


sO X 


o X 


OJ X 


•^ X 


in x 


en x 


o X 


in oj 


o m 


in o 


h* a» 


oo r- 


sO f*l 


^t ^» 


» r- 


CO vO 


-* o 


oj r- 


in 


>o 0^ 


-» OJ 


in o 


m o 


^t CO 






AJ « 



< « 

t- o 
•< o 



r^ x 


CO X 


en x 


OJ X 


CO X 


o X 


in x 


CO *) 


CO (*) 


oo en 


--»o 


vD m 


*-< OJ 


r- ^ 


OJ CO 


CO 


in o 


m r- 


in co 


in m 


»-» OJ 



z 

Q 

z 
tr 

LU 

z 

UJ 

o 

_l 
< 
a 



< 

< 



z 

LU 

u 

-I 
< 
u 

a 

LU 

z 
a 

LU 
I 
H 
3 
O 
in 



a 

LU 



o rvj 



^ X 


^ X 


r- x 


CO X 


-a x 


»-< « 


* X 


O X 


r» x 


-~* X 


O «-i 


oo in 


•-• -* 


in oo 


o oo 


CO CO 


enr- 


OJ O 


co in 


O CO 


oj <r 


v4> r- 


■* >C 


in * 


-* .-« 


^o ^> 


|N T-* 


no ai 


m o 


CO o 



00 

NO 






aj m 



< »» 

h- O 

< O 



(M 00 
OJ O 



00 *e 

r> NO 



o if 



ao o 
r» o 



|v o 

UP -o 



m o 
in --* 



in .-* 
* 1^ 

Lf) OJ 



LfJ (VI 



^f « 


< 


*» 


a »» 


i» »» 


o> Ol 


»- 


<) 


o^ in 


Od D 


(v ir 


< 


c 


C !»■ 


•C 0C 


Cf 




< 


^ rr 


^ o- 









•— 


V 




z 








~o « 


< 


£< 


o « 


* >» 


|v av 


1- 


< 1 


O0 00 


crt o 


(\ oc 


< 


o 


o- r» 


(» .£ 



»-» X 



^ X 


^t X 


m x 


en x 


CO X 


0> X 


so m 


r- r- 


o o 


oj in 


O CO 


r- »c 


a 


cr oj 


OJ ?^ 


^ cr 


r- at 


DC (T 



< X 


OJ X 


OJ X 


»- O 


r- oi 


in -c 


< o 


oO •— 


0J OO 



-r m --i--« 



in x 


r- x 


CO X 


00 X 


co in 


J> O) 


*o |N 


ro --) 


m co 


•-i in 


o> in 


in *-t 



z 





13 




11 


h- 


in 


ii 


M? 


Q 




3 


z. 


on 


at 



D 

z 

Z3 



in 
(1 



X 



5- 
< 
Dr 
13 
CD 
DC 
(1 



li 



uj u. 

o. 



a 



4-66 



xaoana nvioi da iN3DM3d 




u in 
« UJ 










o 


C£ 




_l 




y 


< 




o 


2 


O 




< 




< 


7- 




^> 




LO 




( ) 




rv 


n 




CD 


1- 


■■' 




n 




(H 


.— . 






< 


1 J 




_j 




n 


i/> 




(J) 


tr 


Q_ 








tc 


t/) 




2 z: 


LU 


CO 




CI 




n 


ii i 




i-i a 


D- 




1 


■y 






u 




l- -< 


o 


_l 


l> 


< 




LU 


n 




< _i 


h- 


<. 


tx. 






( ) 


tr 




a: _j 


-J UJ 


cc < 


LU 


10 




ii 




UJ ►-. 


< o 


UJ 


HI 


v- 


n 


> 




/> 


D_ s: 


1- a 


Q tO 


< 


.-- 


tr 


n 


jj 


o 


o r> 


\il 


111 


H 


> 


UJ 


7- 


JJ 


o 


1- CD 


Ll_ 


cc 


u> 


L-L 


to 


< 


u_ 


OJ . 




dagwnN X30NI 
4-67 



CO X 


rn> x 


no x 


ip x 


to X 


<r» cu 


* r^ 


oon 


oo r*. 


„-« *o 


-* oj 


*-* a\ 


wo 


OJ oo 


4- 



-o « 


|v X 


j- 


(\J 00 


CT> -o 


U) 


OJ 00 


rv 





*-* 00 



-4- oo •« 



^ X 


O J* 


IT X 


(T X 


-* «-« 


oo r- 


r- on 


w (T\ 


-* --r 


-M 


ro o 


a\ 



<r ** 


ox 


o 


■^iv 


m oj 


If) 


OJ 00 


v-i 





HI 

J 

< 



.-» m x 



O X 


%*> v* 


r-- x 


ru o 


00 OJ 


>o OJ 


•H ■* 


<£ 


re <£ 



ox 

ao m 



*c X 


OJ X 


OJ X 


o x 


o o-i 


o> 00 


in » 


^ 0^ 


»-» (▼) 


■o 


OJ OJ 


*~* 



un x 


00 x 


r* un 


|v * 


•-• ■* 


0"l 



O X 


«-• X 


to X 


an x 


ox 


O0 sO 


*-t «H 


ao r*- 


m oo 


ro oo 


n o 


— to 


m 


ru --» 


-* 



-o o 



OJX 

iv as 



o 



li 


n\ 


UJ 


•a 


K 


t 


Z 


on 


UJ 


ji 


o 





-1 




< 




o 


uO 



H 


n 


z 


?: 


<?■ 


IT 


LU 

o 


i 


UJ 

X 




<r 


H 


1 


•o 


_> 


< 


o» 


u 


CJ 


■*• 


tfl 



sO X 


on x 


LI) X 


m x 


r*- « 


•jd in 


on OJ 


on co 


ru ru 


--•10 


n ru 


T*1 


OO 


OJ sO 


«n 



<o X 


r-. x 


on j* 


o X 


If) 'O 


O"! OJ 


-« On 


w ^» 


CO oo 


*4 


m 


OJ o 



* X 


on X 


in oj 


>orv 


•MCO 


MH 



X -H X 


IT X 


m x 


IT X 


ru or* 


r* on 


on oj 


on 


C -- ff 


-* 


•r- r- 


r- 



IV X 



iv x 

rH 0C 



ru x 


inx 


< 


■* -* 


soin 


*- 


^ -o 


m 


-* 


• 


• 


(1 


00 


*-* 




vt 


V 




O. X 


<C X 


< 


ro o 


m on 


t- 


«- o. 


r* 


<a 


• 


o 


f"» 


a 


c 




* 


cr 


cr 

z 


in x 


to X 


< 


w r»- 


to OJ 


h- 


•* -* 


IT. 


-* 






K <n« 


00 J* 


OJ X 


|v M 


r») oom 


un o 


fv rv 


in 


j\ * 


rv 


«-< 00 


o 






* X 


OX 


< 


*« 


O OJ 


to r-- 


»~ 


►- 


--i m 


>o 


< 


<t 


• 


• 


o 


t » 


IS 


OJ 






jj 


ro 


tr> 


a_> 



u 






<r 






UJ 












u 




> 


O) 


>- 


t- 




>— 


1 


l_) 


_l 


) 


•-• 


3 


l> 


o> 


u 


< 


< 


< 


li 


on 


u. 


IU 


^ 


UJ 


u 


o> 


3E 


^ 



a 



> j 
< 






s: uj 

— UJ 






— • Q- 



4-68 



siNaanis ivioi jo iN33d3d 





O UJ 


a 


QC Q 


< o 


H 


UJ => 


O 2 




O h- 


_J LU 


1 


2 to 






ZD 


o o 


LO 


UJ 


2 CO >- 


1- 


-1 h- 


•-» H 


-z. 


< < 


IU J 


LU 


O 3 


O i-h ZJ 


o 


i-i Q 


< l/)U 


=) 


O < 


LU < < 


\- 


LU Lt 



LlJ 



i a 


s 




*v% 


N 




*^. 


S 




%.% 


s 






> 






s 






s 




* 






♦ 






« 


w\. 




* 


v» 










» 


\ % 




;> 


\ * 




V 


V \ 




.r 


\ % 




,' 


i v. 




V 


V 




,r 


1 \ 




V 


I \ 




N 


K * 




S 


!\ * 




S 
> 


11 \ 





aaawriN xaasii 
4-69 



o « 

rl O 
00 00 



rl « 

OMfl 

o m 



o ru 
*o in 



in -n 
ru m 



in cn 
oo o 



o « 

Cn * 



rt (\J 
r» 00 



O » 

00 Cn 



OO X 

m ro 



rv x 
in in 

cn 00 



o in 

ro en 



in x 
rv (\j 

f\l r» 



< 

I- 



n o 


■c 


in 


,!* 


* 


X 


or: 


x 


nn 


x 


rr 


x 


h 


» 


^1 


x 


rr 


X 


-f 


x 


<y 


X 


in 


x 


v m 


CII 


OCU 


in 


O 


cn 


rl 


rv 


cu 


in 


< > 


* 


m 


o 


CI 


rv 


cn 


1X1 


hi 


* 


rv 


r* 


< ) 


o r* 


* 


cn *o 


■j) 


-- 


(M 00 


>o en 


rl 


«* 


H 


o 


oo 


m 


O 00 


oo 


* 


iv 


J 


oo 


» 


ri 


in 




■is 


■" 




(VI 


-1 




* 


fn 


ru 




rt 




in 


"■* 


rv 




•« 




in 




■c 


o <o 


x 


-* 


#> 


rv 


x 


in 


X 


<M 


X 


cn 


X 


J 


x 


o 


x 


ru x 


rv 


x 


on 


X 


^n 


x 


•o t~- 


* 


r 


O 


0C 


» 


<c 


o 


rv- 


in 


rv 


IT 


f ' 


rl 


rl 


in 


rl 


C) 


Cn 


< ) 


r> 


cn 


in 


ri 


■■c *c 


f^: 


■< 


a 


* 


|v 


■ H 


•" 


(V 


jj 


rv 


or 


ru 


* 


r»i 


rn 


X 


at 


r^ 


(T 


r- 


-» 


m 


r^ 



O T+ *-* 



i-l r* O 



J) 


ru x 


< X 


o X 


m x 


ro x 


o x 


in x 


ru x 


rl X 


o x 


m x 


rv x 


r- 


-o m 


►- o 


-* « 


m oo 


n cn 


i*i n. 


nm 


-o ru 


cn ru 


o ru 


in <*> 


■o in 


O 


in -o 


< o 


j- <n 


cn ru 


o m 


-o n» 


n cn 


n -a 


so n- 


<o ru 


m oo 


-* -rH 



O 


in 


00 *< 


< x 


O it 


cn j< 


l~- 


in 


cn »-* 


»- o 


t~- cn 


ru ru 





in 


■* ro 


< o 


ru m 


-* o 


CT- 












-j; 


*-i 


^ 


o 


.-i o 


ru •- 



•-• « 


cn 2* 


* in 


m o 


cn * 


■r-iru 



cn *< 
ru w 
r^ rn 



cn cu 

in -< 



ru h 
n* cn 



cn x 

O vO 



z 
•— i 
o 



a: 
in 
i- 
z 
hi 
u 



m 


-* 


o X 


< *« 


j) 


n~ 


ru *-* 


I- o 





o 


in m 


< © 


on 









*o w 


^ m 


cn ;* 


■* X 


ru a-? 


gD >e 


m *e 


o i* 


o o 


OJ CTi 


IT) -r-l 


m ^ 


oo cr. 


r^. m 


n qo 


cn oj 


<r\ o 


m ^* 


■r-« m 


O If) 


* cn 


r^ <* 


-r* CO 


o o 



o .-< *-< 



LU 
O 



< 
CJ 



a 
w 

z 
q: 



CO 

id 

O 



o cn 



< » 
»- o 

t o 

Q • 



oo X 

cn o 

|v o 



in x 

O v0 

cn u» 



m o 
m in 



oo *« 
* o 



o « 
in * 
00 o 

00 



IV « 

in ro 
ru * 



cn j* 
rv m 



m in 
o rv 



oo a 
ru m 

00 00 



cn oo l- o 

-» c < c 
• c 



00 X 


it w 


m t* 


sC X 


M* S* 


-- M 


^» w 


4 cu 


(T. in 


m in 


l»l cn 


o ru 


o ro 


* r- 


oc cr 


0C -- 


c c 


-- c 


a cv 


•- c 


ir ir 






c -- •- 






li 

a 



o 
in 

Ol 



o in cn 
r» ^ oc 



t- o 
< o 



O X 

cn cn 



o * 
cr rv 



ro x 

m m 



oo x 
in cn 



ru x 
o cn 

-- o 



in ru 



o x 

00 .£> 



m cu 

-* o 



< x 
I- o 

< o 
o • 

o 
ib 



o « 

ru oo 
o Iv 



s0 X 

» to 
cn in 



cn x 
cn m 
rv ru 



cn in 
oo * 



CU i< 
■rf o 



(U X 

Cn •-> 
ao m 



in i* 



oo x 
o rv 

in oo 



ru x 

<Q O 



cn x 
m in 



cn *< 
in cu 



tH rl rl LO 



l/> 
o 





CD 




a. 


^ 


in 


o 


«5 


»_1 




3 


S 


ID 


ffj 



L.1 
3 









U 



CO 

UJ 



z 


rr 


UJ 


C3 


5- 


CO 


Z 


or 


Cr 


CL 



< 

li- 



en 
2 



LU _l 

rr ui 

r. (J 



z •— 



4-70 



i39ang nvioi do ±N3Dd3d 




ID If) 






U If) 


















cc o 






tu q: 






Ifl CL 




X-.T 

•:•■ .. ■ 
>:a 
:■:< 
■■•.;; 




x^x 


;■;. ■ 




es 
o 



I 






to 


cr 




_1 




s: < 




Q 


z 


1 ) 




< 




< z: 




r> 




10 




( > 




a: a 




m 


h- 






' i 




O M 






< 


n 




_J 




o to 




CD 


a: 


n 








LC to 




:z z: 


LU 


i/l 




o 




a. lu 




"-" o 


Q_ 




X 


7" 




u_ 




1- -• 


a 


i 


( ) 


< 




LU ("I 




< _j 


h- 


< 


ir 






o cr 




CC _J 


_J UJ 


CK 


< 


LU 


U) 


>-i Q. 




UJ i-. 


< o 


LU 


iii 


h- 


i 1 


> 


.o 


a. s: 


H- Q 


Q LO 


< 


,-' 


cc a 


LU 


o 


O => 


UJ 


LU 


h- 


_-> 


lu z: 


UJ 


N 



s 




hi 


> 




r\ 






> 


? 




M 






/\ 




V 


\ !\ 






i) 




* 


\ 




* 


\ * 




■1 


\ i! 



ubhwdn xboni 
4-71 



* *< 


00 *< 


J> X 


ro .-« 


Ol *< 


0"» *0 


m * 


in o 


^ in 


ai * 


in oo 


ro no 


»-«0J 


m r* 


nO 



•r-» nj -» 



* X 

01 in 



U 

I 



UJ 
J 
ill 
< 



on x 


>o x 


00 rf* 


ox 


-nn 


* cvj 


^-•0^ 


n •-• 


CO (VI 


--ir~ 


m cvj 


(VI 



nj ** 


— • ;* 


OJ >o 


r-. in 


it in 


a in 



O «-* T-* 



fXJ *( 


00 *-- 


<r m 


oj in 


CV fVJ 


rv 



CVJI^ 
O 



ao ao 



in ** 


OJ ** 


(^ ff» 


00 -^ 


-< CVJ 


r** 



* 


X 


m x 


O X 


* X 


ox 


01 


m 


CM CVJ 


f) *-4 


in o 


(VJ Ol 


^ 


a 


CVJ 01 


»-«<o 


CVlvO 


00 



00 ■* 



m x 
00 tn 



•-■ *■<) 



00 X 


I-- X 


in x 


00 X 


o * 


CVJ 00 


ON 


CVJ 4 


CVJ o 


•-. o 


m * 


* 



in at 


m « 


00 ao 


rv a\ 


—Ifv 


*-* 







i- 
z 

UJ 


o^ 


m 


o 


;* 


o x 


•H X 


m *? 


00 X 


z: 




sO 


* 


en 


m 


■^r- 


m o\ 


ojrv 


r*C0 




i 


— * 


* 


-H 


pj m 


-H ^* 


« AJ 


vO 


UJ 




o 


■i 


h 




0C 


oc 


^J 


oc 


.-. 


_J 




-1 


m 






n 


.-• 


•^ 


OJ 




_) 




< 


















o 




u 


















z: 




Q 


o 


en 


r-. 


M 


.-• i* 


00 i* 


*> * 


00 X 




j) 


* 


m 


cm 


m -o 


-H O 


m a 


»-i 00 


or 




s: 


• 


o 


* 


■* 


OJ o 


«-*ra 


fvj u» 




















,_ 






H 




a 






















< 



CV X 


-* X 


O X 


r» X 


4 X 


CVJ C*J 


o ao 


•-• rt 


evien 


w-t cvj 


-* w- 


CV oc 


«- c 


a cv 


4 



TO ** 


in *? 


■JJ 00 


r-. *-* 


-* o 


a> 


• 


• 


CTi 


o 


%C 


rr 


oc »* 


OC d« 


4 00 


^ ^ 


*- f 


* 



I 

«? 



a 
z 



o 


u. 


I— 1 


a 


E 




LU 


V) 


O 


UJ 


< 


IS 


o 


< 


< 


a. 




UJ 




> 



(K J <- — 



CO S* 


in m 


m *< 


o m 


ajco 


■-< d 


«-■ •- 


a a 


c: 



#f m *« 


0J i< 


■v a 


m « 


n o^ oo 


00 ^» 


-nm 


<-« 00 


n <h *-t 


o 


ojr- 


j- 



CT* *e 


^f 


« 


< 


nj ^) 


*c en 


h- 


•H IT 




■3 


*t 


• 




• 


n 


C 




m 




-0 




f) 


ft 


•£ <*■ 


<T\ 


erf 


<» 


--* * 


in in 


h 


H(T) 




o 


< 



I- 

z 



1 


<r 


3 


m 


z 


o 


ii- 


L»J 


UJ 


z 






iii 


1- 


1- 




Z 


z 


i- 


liJ 


liJ 


z 


o 


[11 


tij 


J 


D 



) 
1 1 



U 

_l 



•— O 



) 



or 

CD 



4-72 



siNaanis ~ivj.oi =to iN3oa3d 




■w i - ii -w- ii -wi ii w i -: 




: i/) >- 

I h- 

: c_> _i 

) >-. z> 
: <rt o 
j < < 
- m u_ 



a < 

lu a: 



UJ 
1=1 

•a: 




dBBWON X3C1N1 

4-73 



in « 


o-> « 


ru *< 


-o S* 


o~> « 


>o ** 


r-i *» 


•-I 00 


-« in 


ru * 


in * 


(VJ C\J 


in ct\ 


>o r^ 


J> 4» 


r-v m 


o * 


-» in 


r* -h 


•o o\ 


m 



no x 


r^ a* 


00 i< 


00 .-« 


in co 


ru r-v 


o in 


ao »H 


<M o 


* o 


* vO 


* -* 



< 

CM 



I 
< 



— X 


r^ x 


-< X 


00 ^ 


* x 


o X 


O X 


o x 


-* x 


(T X 


(? X 


01 00 


CO t-» 


do ro 


r- »-• 


ru r*- 


oi in 


O 00 


*o O 


■^ r»- 


nO o 


ru ro 


-* ■* 


-* yO 


(VI o 


ru * 


m o 


* 


C\J o 


in oi 


m o 


o^* 


ru ru 



^ r» x 
n o ^t 


< x 

1- o 

< o 
o 

CD 

z 


01 x 


NO X 

<c ro 
r^ a 


m a> 


OO X 


o X 


r- x 


* x 
ac ru 
m o 


ro ** 
<X) ac 


cu X 
(V o 


in ru 

O -r- 


■v ai 


- * «-l 


ru -vD 


m 


in ru 

ro 


o 


.-» sO 


r-* 00 


.-• oo 


m 


--• \0 


t> oi x 

n n-c 
* mo 


< a* 
►- o 

< o 


r^ x 

--t 00 


in x 

00 * 


r*** x 

-* ro 
vo en 


in x 
oi ro 


in x 

ru ^ 


CO X 

r- in 

o m 


o in 
in -h 


00 X 

0^ 01 

ro -* 


r- x 

r^ -jo 
m ro 


-* in 
oi rv 



o X 


< X 


O X 


in x 


r^ x 


O X 


01 X 


*> in 


1- o 


ro * 


ru in 


01 ^» 


ru 4 


ru i^- 


ro «-» 


«* o 


■^ r- 


co oa 


o in 


OJ *-< 


«-• r^ 



■^ »h in * 



oo X 
so m 



i «h r- m 



o 

z: 



T) »-H X 


< X 


«-* X 


in x 


o X 


in x 


■* X 


O X 


»-• X 


rn x 


■rn X 


01 X 


o o * 


1- o 


^« ^ 


ro ^o 


m oi 


oi m 


ro ru 


^o ru 


in r^ 


--• in 


oo 


sO o 


■»» mm 


< © 


ru in 


-4) ^4 


DO O 


vfi 01 


t-* 00 


ru rv 


ru x 


m o 


^* -* 


-o «■-» 



IX 




n 


LU 




hi 


h- 




i 


^ 






111 


•J 


/ 


c> 


h- 


ii 




01 


in 


1 


rt 


K 


<r 


1 


1/) 


< > 


■J 


III 


(—. 


M.1 


5 


a 


a> 




UJ 


■H 


O 



m x 


< X 


-* X 


01 X 


rn x 


ru x 


01 X 


■^-t X 


—i s0 


t- o 


sO ^t 


in .-« 


-t ao 


cu ^» 


o --» 


•43 w 


m o 


< o 


-* ru 


* ru 


o -o 


<r> o 


--« r*- 


ru m 



ru x 


on 


,^ 


ru ru 


at 


ai 


00 ■* 


* 


'VJ 



< »» 


at « 


a »» 


IT »« 


IT •« 


•- « 


OC « 


f^ M 


IT »« 


* » 


r^ tr 


h- O 


rvj in 


fn o> 


m r»i 


m ^ 


X <c 


a in 


m o> 


^O 1^. 


en oo 


O OC 


< C 


r^ ir 


p- a 


cr n" 


(V c 


*- (V 


*- or. 


a 


rr ir 


^ or 


ir o- 



Q 


UJ 


< 


o 


O 


< 


< 


rr 



^ X 


< 


^ 


J- X 


fv x 


O X 


m m 


>- 


o 


ru -^ 


m oo 


^* in 


rr a- 


« 


o 


ir r» 


rv lt 


oc r*i 



o « 


m 


<it 


rvi « 


in « 


rvi r^ 


r^ 


» 


OJ <T\ 


vo r^ 


— 


« 


-» 


4 sC 


4 o 



*H H G*\ 



•* X 


OJ X 


O X 


01 X 


00 X 


O X 


0J X 


r- X 


ro x 


r^- x 


00 X 


K O 


ru in 


* in 


--• in 


OJ o 


o m 


oi m 


r- 


^» ^ 


00 00 


01 OJ 


< O 


m m 


o o 


ao O 


* m 


H 01 


oo ro 


<1 


r^ oi 


in * 


o in 


O • 


^ • 


^ • 


• 


•* • 


* 


• 


• 


• 


• 


• 



.2 



13 

i J 

=> 
m 

in 

z 



1 1 
) 



1 

I) 



rr 



i/> — • 



4-74 



i33ang nvioi do ±N3:>u3d 




n 


■ n 




I/-. 


11 






ii 


III 




u 


it 






■> 




- 




i" 




to 


•: 







l-H 


i/) 


O 




a: o 






h- 


2T 


n 




(J> KH 






< 


a 


_i 




O CO 




LD 


CE 












UJ 




o 










D_ 


i 






u_ 




1- -< 
< -I 

or _i 


O 


—1 o 


< 








1- 


< tr 






o a: 




-1 UJ 


a: < 


UJ 








< o 


UJ UJ 


1- 


n 


> 


i/l 


a. s: 
o 


h- Q 


Q LO 


< 


?" 


or q 


UJ 


□ =3 


UJ UJ 


1- 


-» 




UJ 


1- m 


u_ q; 


l/l 


u_ 


CO < 


U_ 


^ 




d3BWnN X30NI 

4-75 



( > 


*-t it 


in » 


r^ ** 


* J< 


r-t * 


tvj 


>o >o 


00 (\J 


o * 


Jt -o 


-o oj 


in 


♦ m 


n (*i 


-t o 


in r- 


o 



01 no 

r\j o~* 



ro « 

o .h 
•-• o 



nC » 


00 J* 


OP W 


in « 


(Ti a* 


PO «-« 


.-♦in 


or- 


o\nj 


* in 


-* *H 


-* 00 


.M ■r-t 


^* cn 


rvj 



r* so 



UJ 

_l 

OJ 

< 



--* *e 


* at 


en *< 


O i* 


OO if 


.-. 01 


C 00 


00 OJ 


in r- 


& fr. 


-» o 


* ir. 


r^ 


J» o 


<C 



-h cn « 



o w 
(n in 

m ro 



m pj 
pj 



o^ oe 



■r-t « 

f*> o 



r- >« 


-H « 


m o 


oi en 


oj ro 


-o 



\J -* « 


r- *? 


ro « 


oo «-e 


PJ i* 


O ji ("vi 


%© m 


om 


-h pj 


r- in 


-« mm 


OJ o 


■r-t 00 


-* o 


CU 



oo S< 


OJ a* 


*< 


^-* m 


00 vO 


I- 


p« in 


^t 


< 


• 


• 


a 



ao rf* 


T-t « 


-r-t *« 


-* x 


00 »■? 


ro -* 


G\ 00 


a\ r- 


a*oj 


-■*■ o\ 


ro -* 


*-«PJ 


OJ 


rn in 


ro 



oo PJ 
~* oo 



OJ ** 


r- s* 


o « 


00 ^ 


ao *« 


m w-t 


or- 


00 r* 


tn r- 


OJ rH 


ro o 


pj in 


o 


en in 


00 







n 


it 




in 


Ill 


■* 


> 


h- 


r- 




^ 


tji 


z 


UJ 


~H 


It 


O 


1 


III 




■* 


K 


—I 


tf> 


w 


< 


(T- 


in 


L) 


~ 


* 






in s< 

00 0> 



Q 
< 

< 



O <VI * 

m run 



a 

a 



IT. K 


at M 


0- K 


4» »f 


^ w 


ou rs 


*> m 


oo r- 


oj m 


ru r- 


pr r- 


^ c 


if 


o* r- 


oc 



oj o> 

■c 



in 

4! 



4- « 


00 rf< 


OJ ** 


■r* *< 


r- « 


^t %c 


-o o 


r- r- 


m CTi 


PJ ^o 


ro m 


--< * 


0^ 


m pj 


en 



in r- 


r- 


UJ 


»- 


► 


-- r- 




cv 


-a 


< 


• 




• 


C 


r* 


of 




~~ 






^ 




f. 


a 


a 


on i* 


00 


*e 


-* 


< 


m yf) 


4) 


m 


i- 


1- 


»- o 




ir 


•4 


-a 


• 




* 


O 


n 


r*. 




iu 






43 




r-n 


or 


07J 

z 


^-i *f 


00 


^ 


<t 


<I 


m o> 


U! 


o 


k 


1- 






U) 

nr 



3 u 



2" 


11 


lil 


> 


(1 


>- 


> 


U) 


1- 




U] 


i 




•i 


1 


i ) 


•J 


•— • 


1- 


'i 



•-• w 



111 


n 


1. 


..i 


•-• 


ul 


1— 


K 


1 


^ 


K 


~» 


a 


1 


< 


J) 


a 


> 



4-76 



siNaanis ivioi jd iN3D«3d 









U38WnN X30NI 

4-77 



NEW ACADEMIC MEDICAL CENTER FUNDING TABLE 4A-13A 

10 Schools* 1967-1974 
Totals in Thousands of 1964 Constant Dollars 



Year of Operation 



Total Operating 

Budget 36,763 45,987 59,692 77,569 89,974 96,905 

State and Local 13,429 15,582 19,146 20,126 24,886 29,910 

Government Funds 37% 34% 32% 26% 28% 31% 

Federal Research 9,865 13,615 14,069 16,518 19,674 20,932 

Funds 27% 30% 24% 21% 22% 22% 

Service Programs 2,110 2,882 4,413 16,608. 17,459 24,086 

& Professional Fees 6% 6% 7% 21% 19% 25% 

Capitation and 695 1,045 1,609 2,695 3,591 4,718 

Special Projects 2% 2% 3% 3% 4% 5% 

Federal Teaching 1,585 2,212 3,084 5,370 4,549 5,576 

and Training 4% 5% 5% 7% 5% 6% 

Non-Federal 624 851 1,163 342 4,152 1,928 

Research 2% 2% 2% 0% 5% 2% 

Endowments and 306 1,141 767 1,489 1,946 3,015 

Gifts 1% 2% 1% 2% 2% 3% 

Other 8,149 8,659 15,441 14,421 13,717 6,740 

22% 19% 26% 19% 15% 7% 



^Synchronized by year of operation rather than by calendar year. 



4-71 



FIGURE 4A-13B 



NEW ACADEMIC MEDICAL CENTER FUNDING 



AVERAGE OF 10 CENTERS 



10 



to 

tr 
< 

o 
a 

z 6 
< 



o 
o 

10 
0> 

Li. 

o 

(/) 

z 
o 



4 - 



2 - 



TOTAL OPERATING 




PRO. FEES 




NON-FEDERAL 



-STATE & LOCAL- 




-80 



x 
o 
m 
z 
-60 H 

o 
■n 



100 



"40 



w 
c 
o 



20 



12 3 4 5 6 



I 2 3 4 5 6 



YEAR OF OPERATION 



u 

ro 



I 

M 
D3 
EH 



^ 




Eh 




>J 




D 




U 




< 








Eh 




s 


■3" 


w 


r> 


s 


en 


j 


H 


H^ 


I 


O 


r- 


Pd 


>x. 


a 


Oi 


w 


rH 


tf 


* - 


w 


0) 


Eh 


H 


a 





w 





u 


£ 




o 


1-1 


w 


5! 




u 


o 


H 


H 


Q 




H 


SH 


§ 







MH 


u 




H 


to 


s 


rH 


w 


rcS 


Q 


4-1 


Si 





U 


Eh 


< 




S 




w 




^ 





o 

•H 

-P 

rd 

Sh 
(1) 

O 

4-1 

o 

M 
rrj 

tu 



ml 



■*-! 



o 


co o\° 


O o\° 


co o\° 


O o\° 


O o\° 


CN 


rH CM 


CM T 


LO r-l 


co ro 


O ro 


r>- 


<h< ^ 


CO iH 


LO rH 


CO CO 


CN 



ro 


rf 0\° 


ro 


0,0 


^r 


01° 


LD o\° 


CTi o\° 


CO 


rH rH 


CO 


ro 


m 


rH 


O CO 


CM CO 


CTi 


o "3" 


LO 


rH 


in 


rH 


LO CO 


rH 


N 


*. 










*. 




vr 


CN 










H 




rH 


CM o\° 


t- 


0\° 


CO 


<#P 


CO o\° 


rH o\o 


CO 


LO CN 


CO 


«tf 


rH 


CN 


r» o> 


•>* CO 


■^r 


CO "* 


LD 


rH 


LT) 


rH 


CM CM 


rH 



r~- 


LO o\° 


CN crP 


O tic 


00 o\° 


rH 0\° 


IT) 


LO CTl 


CXI CO 


CN ^ 


O r-{ 


CO CO 


LTt 


CO CN 


LO CM 


T rH 


CTi CO 





CO 


<J\ dP 


IT) o\° 


CO o\° 


LD o\° 


rH C*° 


LO 


r~ co 


LO CM 


O CO 


O LO 


o r- 


1 ro 


CO CN 


r-\ rH 


rH 


LO <^ 


rH 





to 




4-) 


CD 


a 


4-> 


Q) 


C 


T3 


0) 


3 


■a 


4-> 


3 


CO 


4-> 




W 


H 




CO 


rH 





rd 


•H 


4-1 


T3 





(1) 


E-< 


g 



4-1 










c 








03 a) 


0) 








rH U 


T3 03 








rrj C 


P 4J 






T3 


Sh 03 


4> C 






c to 


O -H 


CO 




0) 


m 4-i 


4J U 


rH 


a) 


4-> 


c 


U CO 


H n3 


4-1 


C 


ra <d 


o 


rd > 


fO 


0) 


C T3 


a u 


U -H 


3 nrj 


Sh -H 


1 -H 


•rH 3 


T) 


3 


CD 0) 


4-> 03 


TJ fJ 1 


rrj 


4-> 


4-> a) 


CD rO 


a) w 


r4 


CO 


c a 


CQ 


S 


rj 




H 


04 













H 












rrj 












0) 












>i 


o 


*-r 


o¥> 


VO 


&f> 


lH 

rrj 


rH 


CO 


LD 


CM 


LO 


CO 


r-\ 


LD 


VjO 


CO 


*• 


- 








Q) 


<-t 


rH 








rH 

rd 



*-\ 


CN 


o\° 


o> 


o\o 


X) 


in 


O 


LD 


•^ 


ID 


LO 


O 


LO 


LO 


CO 


rH 


H 








4-1 


LD 


in 


0\° 


T-\ 


oY> 


H 

a; 

d 


>* 


LO 


CN 


00 


00 


LCI 


CTi 


l£> 


in 


CO 



(N 


rH crp 


ID <A° 


^T CrP 


r- ov> 


CO df 


<y\ 


LO dp 


■* oV 


Lfl 


r^ ix> 


CTl <3< 


o ^ 


o ^ 


LD ^J- 


CM 


rH rH 


rH a\ 


LO 


CM ro 


^f rH 


LO rH 


CM CO 


rH 


CO 


CO VD 


in co 



CO 
(T. 



LO 



(U 
g 

•H 

EH 

I 



3 
fc >1 

4-1 
~rH 

rH 3 

rrj U 
4-) rrj 
(h 



^T OP 
rH CN 



LO o\° 
CO CTl 
CO LO 



0) 

u 
c 

0) 

H 

u 

co 



U 



4-1 

rH 

u 

C3 



CTl cr\o 

r-- co 
co co 



rH o¥> 
(^ rH 
CN •<* 



V 

c 

0) 
•H 

u 

co 



>1 

4-1 



U 3 

-H U 

U) id 

rrj Cn 

cq 



4-> 

rfl 
M 

fl 


■H 
4-1 
rrj 
U 
0) 

a. 

o 

4H 


Sh 
rrj 
01 
>i 

>i 
XI 

13 
OJ 
N 
■H 
C 

Sh 

u 
c 
>l 

CO 

* 



4-80 



siNaanis -ivio± jo iN33d3d 







l±J ZD 



•-<=> 


o 


i-< Q 


t/io 


z> 


Q < 


<< 


H 


lxj a: 


mil. 


l/) 


s: o 



1=1 




31d03d dO S30wnN 

4-81 



B. SPECIAL STUDIES 

1 . Construction 

Data have been obtained to ascertain what part of funding of 
new construction and renovation has been provided by federal 
sources. Table 4B-1 presents this data which was gathered from 
reports of the academic medical centers to AAMC. The total number 
of institutions reporting construction has declined slightly since 
fiscal year 1969, the first for which reliable data is available. 
About two-thirds of the institutions reporting construction re- 
ceived federal assistance. These institutions received about 40% 
of their construction funds for biomedical and behavioral research 
facilities from federal sources even though up to 801 is allowed 
by law. Some of the funds reported were used for the building of 
classrooms and other educational facilities under the auspices of 
the Bureau of Health Manpower. We have been unable to separate 
funding for these latter facilities from those expressly for re- 
search purposes. Nevertheless, it appears that federal construc- 
tion support overall has averaged about 50 million dollars annually; 
this has permitted the construction of buildings valued approxi- 
mately two-and-a-half times more. 

According to a survey of business officers at a sample of 
academic medical centers, once buildings are constructed and equip- 
ment purchased with federal funds, all academic medical centers, 
whether public or private, treat these facilities in the same 
manner as if they had been constructed with non-federal funds. 
No special arrangements have been made for maintenance of buildings 
or equipment which are different than those for the centers over- 
all, except in rare situations. 



4-83 



TABLE 4B-1 
1 



FEDERAL CONTRIBUTION TO ACADEMIC MEDICAL CENTER CONSTRUCTION 



Fiscal Year 
1969 1970 1971 1972 1973 1974 

No. of Institutions 
Reporting Construction 45 49 42 37 35 37 

Total Cost of Construc- 
tion Completed 2 $135. 8 3 158.1 129.1 155.5 97.8 211.5 

No. of Institutions 
Reporting Federal 
Construction Support 29 34 31 23 22 20 

Total Cost of Con- 
struction Completed 
with Some Federal 
Support $120.9 138.1 108.5 135.9 82.1 181.6 

Amount of Federal 
Construction -. 

Support $ 47.7 47.8 45.6 50.2 40.5 71.1 

Percent From 
Federal Sources 39.4% 34.6 42.0 36.9 49.3 39.0 



Source: Annual Questionnaire of the Liaison Committee on Medical 
Education, Part II. 

2 
Includes only construction completed in a given year and excludes 

construction initiated and construction planned. 

Dollar amounts are in millions of 1964 constant dollars computed 
with Halstead's Construction Price Index. 



4-84 



2 . Organizational Complexity in Academic Medical Centers 
A special study was undertaken to examine the question whether 
federal research funding was related to increasing organizational 
complexity in the nation's academic medical centers. The medical 
centers were ranked in quartiles according to the degree of their 
research involvement (as described previously in the Methods sec- 
tion) . Using data in AAMC files as well as telephone interviews 
with some administrative staff of the centers, the number of health 
profession schools, affiliated hospitals and health science "cen- 
ters" were determined. Health profession schools included schools 
of dentistry, nursing, pharmacy, public health and allied health. 
An affiliated hospital was enumerated only if a significant part 
of the teaching and/or research of the academic center was report- 
ed on the center's annual report to AAMC. The presence of any or- 
ganized research and development center (e.g., primate center, 
child development center, special center of research, comprehen- 
sive cancer center) was recorded and a number indicating the total 
complexity of the academic medical was thus derived. Table 4B-2 
presents the data obtained from all 86 medical centers ranked in 
quartiles by their research involvement. Analysis of variance was 
undertaken; a highly significant relationship was found between 
academic medical centers in the first quartile for research involve- 
ment and the number of research and development centers established 
at those institutions. There was no other discernible difference 
between institutions with regard to the variables investigated. 
Specifically, the increase in federal research funding appears to 
be related only to the establishment of research and development 
centers and not to the number or kind of health profession schools 
or the number of affiliated hospitals in the academic medical 

4-85 



TABLE 4B-2 



RESEARCH INVOLVEMENT AND ORGANIZATIONAL COMPLEXITY 



Grouping By 
Federal 
Research , 

Involvement 



ORGANIZATIONAL COMPLEXITY 
AS MEASURED BY NUMBER OF: 

Research & Health 
Development Profession Affiliated 
Centers Schools Hospitals 



1st Quartile 
2nd Quartile 
3rd Quartile 
4th Quartile 



129 



68 



57 



6 9 



21 



31 



41 



39 



154 



171 



129 



159 



Variance Ratio 



6.00* 



1.39' 



1.22- 



See Methods Section for categorization of 86 schools by research 
involvement. 

"Significant at 5 percent level. 

Not significant. 



4-86 



centers. It should be noted, however, that medical centers were 
extremely variable in degree of total complexity and in their 
arrangements for the administration for these centers. Neverthe- 
less, the magnitude of funds for the support of research did not 
appear to be related to the overall complexity of these institu- 
tions . 

3 . Curriculum Change and Federal Research Funding 

The purpose of this special study was to discover whether 
there were any effects of research funding on educational programs 
in academic medical centers. Specifically addressed was the ques- 
tion: In what ways has NIH research funding of centers, other re- 
search and research training affected the medical school curricu- 
lum in recipient schools? Data was assembled from handbooks and 
files maintained by AAMC relating to curricula at the medical 
schools. Personal interviews were also used to verify data obtain- 
ed from these files and, particularly, to ascertain when various 
programs were instituted and with what impetus. 

The assumption was made that all schools prior to 1965 (with 
rare exceptions) used a Flexnerian system of departmental basic 
science instruction for two years followed by clinical departmen- 
tal teaching in the last two years. Based on this assumption, it 
was found that of the 84 schools reporting, 65 had undergone a sig 
nificant curriculum change since 1965. We were further able to 
ascertain that 34 of 43 publicly-owned institutions have under- 
gone a significant change from the older, Flexnerian curriculum. 

A specific study addressed the question whether federal re- 
search involvement was related to the presence or absence of 
curriculum change. Data to answer this question is presented in 
Table 4B-3A. There appears to be a tendency toward but not a 

4-87 



statistically significant relationship between the amount of fed- 
eral research support and curriculum change. The availability of 
other federal funds (e.g., special project grants) does not clear- 
ly correlate with curriculum change, the distribution being quite 
similar to that seen in the overall breakdown of schools cited 
above (4B-3A). 

It was suggested that the presence of a medical scientist or 
formal concurrent M.D./Ph.D. degree program might be directly re- 
lated to the availability of federal research funds and a prepon- 
derance of research activity in a given institution. Analysis of 
the data presented in Table 4B-3B indicates that greater research 
involvement was associated with a larger number of formal M.D./ 
Ph.D. programs, many of which received direct NIH support from 
the National Institute of General Medical Sciences Medical Scien- 
tist Program. These awards tended to be made to research intens- 
ive institutions. However, a greater association is observed be- 
tween concurrent M.D./Ph.D. programs and private institutions as 
opposed to publicly owned schools. Private schools also received 
13 of the 15 NIH grants awarded for this program. In addition to 
the 15 awards made by NIGMS, an additional 14 concurrent degree 
programs have been established with local funds and were operation- 
al in 1974. These programs were distributed throughout the remain- 
ing three research quartiles of institutions. 

A further question relating to research funding and curricu- 
lum change concerned the initiation of family medicine programs. 
Such programs have been started in the majority of academic medi- 
cal centers within the last five years. As is seen in Table 4B- 
3C, the preponderance of the programs are in public institutions 
as opposed to private institutions (chi-square = 8.37, significant 

4-88 



TABLE 4B-3A 
FEDERAL RESEARCH INVOLVEMENT AND CURRICULUM CHANGE 

1973-74 



Grouping by _ . . „, 

Federal Research! Curriculum Change 

Involvement Yes No 

2 
1st Quartile 19 2 



2nd Quartile 16 4 

3rd Quartile 15 6 

4th Quartile 15 7 

.3 



Chi-square = 3.62' 



-*-See Methods Section for categorization of 86 schools by 
federal research involvement. 

2 

Number of Schools. 

3 
Not significant at 5 percent level. 



4-89 



TABLE 4B-3B 



MEDICAL SCIENTIST (M.D.-Ph.D.) PROGRAMS 
AT ACADEMIC MEDICAL CENTERS 



All Centers 
Public Centers 
Private Centers 



M.D.-Ph.D. PROGRAM: 

Yes No NIH Supported 

29 43 15 

16 26 2 

13 17 13 



Chi-square = 0.04- 



Grouping by Federal » 
Research Involvement: 



1st Quartile 




11 


6 


2nd Quartile 




5 


11 


3rd Quartile 




9 


12 


4th Quartile 




8 


14 




Chi-square = 


= 4.55 3 





11 

4 





Not significant at 5 percent level. 

I 

'See Methods Section for categorization of medical centers by 

federal research involvement. 

Not significant at 5 percent level. 



4-90 



TABLE 4B-3C 



FAMILY MEDICINE PROGRAMS 
1973-74 



MEDICAL SCHOOL OWNERSHIP 



Public Schools 
Private Schools 



Chi-square = 8.37 



2. FEDERAL RESEARCH INVOLVEMENT 



Family Medicine 
Program 


Yes 




No 


34 




4 


18 




14 



Federal 
Researc 
Involvement^ Program 



Research Family Medicine 



Yes No 

1st Quartile 12 5 

2nd Quartile 13 5 

3rd Quartile 14 4 

4th Quartile 13 4 

3 
Chi-square = 0.32 



Significant at 5 percent level. 

2 
See Methods Section for categorization of schools by federal 

research involvement. 

3 
Not significant at 5 percent level. 



4-91 



TABLE 4B-3D 



PRESENCE OF ACCELERATED PROGRAMS 
1973-74 



MEDICAL SCHOOL OWNERSHIP 

Public Schools 
Private Schools 



ACCELERATED PROGRAM 



Yes 


No 


14l 


29 


9 


32 



Chi-square = 0.71^ 
2. RESEARCH INVOLVEMENT 

1st Quartile 
2nd Quartile 
3rd Quartile 
4th Quartile 

Chi-square = 0.52 2 



ACCELERATED PROGRAM 



Yes 


No 


6 


15 


5 


16 


6 


15 


6 


16 



3. SPECIAL PROJECT GRANT SUPPORT ACCELERATED PROGRAM 



Yes 
No 



Yes 


No 


18 


24 


5 


30 



Chi square = 6.14" 



Number of schools, 



Not significant at 5 percent level. 

3 . 
Significant at 5 percent level. 



4-92 



at the 5% level) . Of considerable interest was the finding that 
biomedical research involvement was not related either positively 
or negatively to the presence of a family medicine program (Table 
4B-3C(2). This finding runs counter to the opinion frequently 
voiced that research intensive institutions are less likely to ini- 
tiate programs in family medicine and/or primary care. 

Finally, the association of research with the presence or ab- 
sence of accelerated programs for medical education was investi- 
gated. Like family medicine programs, shortening of the medical 
curriculum was promoted by a combination of federal and state ini- 
tiatives for educational change within the last decade. Conse- 
quently, it is not surprising that public schools tended to have 
more accelerated programs (although the difference was not signi- 
ficant at the 5% level). Further, those medical schools to whom 
federal special project grant support was awarded showed a prepon- 
derance of abbreviated medical education programs. There was, how- 
ever, no relationship between research involvement and the presence 
or absence of an accelerated program. 



4-93 



4 . Changes in Distribution of NIH Awards 

A special study was undertaken to discover what changes have 
occurred in total awards to institutions from NIH over the period 
for which data is available from the IMPAC files (1968-74). Fig- 
ure 4B-4 indicates the number of schools receiving awards for fed- 
eral research in each of 12 $3 million intervals. In 1974, as 
compared to 1968, more institutions received a larger number of 
research funds. However, the distribution of funds was more even, 
than in 1968. Further analysis will be needed to discover whether 
this trend will persist or is transitory. 

5 . Time Serial Changes in Academic Medical Center Funding 

In the analysis of trends in funding of academic medical cen- 
ters it is not possible to ascertain causality using simple corre- 
lation techniques. However, by the use of cross correlation tech- 
niques available with computer analysis we may describe the time- 
dependent relation between two variables. Thus, if variable A 
increases, we may examine whether variables B, C, etc., change, in 
what direction and at what time intervals. Figure 4B-5 displays 
such relations between federal research funding and other variables 
taken one at a time over a time lag of up to 3 years. For example, 
in Figure 4B-5(a) an increase in federal research funding is nega- 
tively correlated with non-federal funds for the first year. Then, 
at 2 and 3 years, there is a positive correlation (increase). Con- 
versely, a decrease is followed in 2 years by a decrease in non- 
federal research funds. Again, causality cannot be assumed but 
the existence of a "delayed multiplier effect" can be documented 
at least in part. The Rand Corporation has estimated the magnitude 

of these multiplier effects in Task 3. 



4-94 



CHANGE IN DISTRIBUTION OF NIH AWARDS AMONG ACADEMIC MEDICAL CENTERS 

1968-1974 



1 



FIGURE 4B-4 



1968 N = 99 




18 24 30 



40 -l 



30 - 



20 - 



10 - 




1974 N = 101 



fj 



ft 




fy///f//?/s/fjy//// r „ys/j 



18 24 30 36 



I 



SOURCE: NIH/lMPAC FILE. 



MILLIONS OF DOLLARS IN NIH AWARDS 



4-95 



FIGURE 4B-5 

TIME VARIATION OF ENROLLMENT/EMPLOYMENT IN 86 
MEDICAL SCHOOLS WITH FEDERAL RESEARCH FUNDING 

1964-74; 1964 CONSTANT DOLLARS 



Cross-Correlation of 
Federal Research 
Funding with: 



a. Non-Federal Research 



Time Lag in Years 
12 3 



""TV- 





L^. 



+ 0.5 
-0.0 

I 0.5 



+ 0.5 



b. Medical Student 
Enrollment 






0.0 

■0.5 



c. No. of Clinical 
Science Faculty 



fj 



'//A Vj 




MMMJt 



+ 0.5 
0.0 

-0.5 



d. Graduate Student 
Enrollment 



vzz 



'Mm 



+ 0.5 

0.0 

-0.5 



4-96 



Examining the time serial changes presented, we find the 
strongest "attractive" correlations between federal research fund- 
ing and a) non-federal research funding, and f) graduate student 
enrollment. Negative multipliers were most evident between fed- 
eral research funding and e) medical student enrollment and, to a 
lesser extent, g) size of clinical faculty and d) service programs 
and professional fees. Correlations were significant only in the 
case of federal and non-federal research funding (a) . 

6 . Patterns of Expenditures 

In fiscal 1972 100 academic medical centers had data suffi- 
cient for analysis of their budgetary expenditure patterns. In 
1972 their total operating budgets were 1.29 billion dollars (aver- 
age 12.9 million, range 1.1 to 43.3 million). Six funding patterns 
were recognizable when revenues for the years 1964-74 were graphic- 
ally displayed: 

Number of Average 1972 
Type Medical Centers Total Budget Range 

1. Stable 9 10.3 million 1.1-22.4 

2. Stable, then rising 10 11.2 4.2-26.0 

3. Rising, then stable 18 16.5 4.1-40.4 

4. Rising at moderate rate 40 14.7 3.2-43.3 

5. Rising at rapid rate 20 8.6 2.0-24.6 

6. Declining or irregular 3 6.5 3.7- 9.1 

7. Administrative Arrangements for Biomedical Research 

By 1974 25 of 116 medical centers had established identifiable 
administrative units created to manage biomedical research funds. 
An additional but unknown number of centers had incorporated simi- 
lar functions into existing offices. Most of the identifiable 
units are headed by associate deans who review grant and contract 

4-97 



applications for commitment of resources and other institutional 
concerns, develop and manage resources for research; act as "gate 
keepers" of research opportunity information; allocate general 
research grant funds, space and equipment; assure institutional 
compliance with federal regulations, and interact with other local 
and national units responsible for biomedical research. All ad- 
ministrative units were formed to improve grant/contract admini- 
stration; most also seek to develop new sources or increased re- 
search funding. 

8 . Trends in Total Federal Revenues to Academic Medical 
Centers 

Biomedical research funds comprise an important source of 
revenue to academic medical centers. As has been noted previously 
(Tables 4A-1F and -1G), the proportion of academic medical center 
total operating budget provided by this source declined from 36 
to 21 percent between 1964 and 1974 although actual dollar reve- 
nues increased. We have examined other federal revenues and find 
(see Table 4B-6) that federal teaching and training revenues de- 
clined slightly while federal capitation and special project 
grant support increased as a percentage of total operating budget. 
We have been unable to estimate other possible federal revenues 
to academic medical centers (e.g., through science programs or 
Medicare reimbursements); however, it appears that total federal 
revenues to medical centers have not been rising as fast as other 
revenue sources. The result is a decline in percentage of total 
federal support. 

9 . Career Choice of Medical Graduates 

An oft-quoted impression about medical education is that the 
national commitment to biomedical research has served as an impedi- 
ment to the production of doctors who take care of people. On 

4-98 



TABLE 4B-6 
PERCENT OF TOTAL ACADEMIC MEDICAL CENTER OPERATING BUDGETS FROM FEDERAL SOURCES 

Fiscal Year 
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 



Federal Research 

Direct Costs 364 35 34 33 29 25 21 22 22 21 

Indirect Costs 6767766676 

Total 42 42 40 40 36 31 27 28 29 27 

Federal Teaching 

and Training 12 13 13 13 12 11 11 11 11 10 

Federal Capitation - - - - 1 2 2 3 3* 4 6 5 

Total Federal 54 55 54 55 50 45 41 43 46 42 



4-99 



the basis of available data on the career choices of medical grad- 
uates, this does not appear to be the case. Table 4B-7 shows the 
comparative numbers of medical graduates entering patient care vs. 
research and teaching from high research-intensive and low research- 
intensive schools. An important result is that the numbers of 
graduates entering patient care from high-research and low-research 
schools (87 percent and 94 percent, respectively) are not signifi- 
cantly different. The corresponding numbers of graduates entering 
teaching and research on the other hand, are significantly dif- 
ferent at the 5 percent level, but they account for only 10 per- 
cent of the total graduates of research-oriented schools and 4 per- 
cent of the graduates of less research-oriented schools. On the 
basis of these data, research-oriented academic medical centers 
appear to be just as likely as less -research oriented centers to 
turn out graduates who choose to go into patient care. 



4-100 



TABLE 4B-7 

COMPARATIVE NUMBER OF MEDICAL GRADUATES ENTERING 
PATIENT CARE VS. RESEARCH AND TEACHING 1 

1965-69 



Group 

1st Research 
Quartile^ 



4th Research 
Quartile^ 



Student's "t" Statistic 



Number of 

Total 
Graduates 


Numb e r 
Entering 
Patient 
Care 


Numb e r 
Entering 
Research and 
Teaching^ 


8,105 


7,023 c 
86. 7% 5 


841 
10.4% 


8,636 


8,159 

94.5% 


316 

5.7% 


tistic 


1.02 6 


4.81 7 



Data from Medical School Alumni (16) 

2 

21 high research- intensive medical centers. See Methods 

Section. 

22 low research-intensive medical centers. See Methods 
Section. 

Graduates entering patient care, research and teaching do 
not account for all medical graduates. Graduates entering 
administration and other professional activities are not 
shown here. 

Percent of total. 

Not significant at 5% level (41 degrees of freedom). 

7 

Significant at 5% level (41 degrees of freedom). 



4-101 



5. DISCUSSION 

Our task has been to collect and present data which will de- 
pict the dimensions and trends in funding of academic medical cen- 
ters between the years 1964 and 1974. Our further task is to 
assess the impact of biomedical and behavioral research funding 
from the National Institutes of Health, the National Institute of 
Mental Health and related agencies on academic medical centers 
over the same period of time. We have attempted to describe the 
trends and impact from the standpoint of the academic medical cen- 
ter but it has not been possible to describe these changes quanti- 
tatively with as much exactness as was initially hoped. Further, 
it has been very difficult to depict the subtle and often quali- 
tative changes which have been the result of changing goals, bio- 
medical support mechanisms and programs . Before turning to the 
appraisal of the results of this study, it is useful to examine 
the recent history of biomedical research from the point of view 
of the funding agency, principally NIH. In making this assessment 
we have been very greatly assisted by Dr. James Shannon, former 
director of the National Institutes of Health, who has made an ex- 
tensive review of the postwar history of NIH (8) . 

For the most part, the roles of NIH/NIMH cannot be adequately 
assessed by viewing only the last ten years. For the clearest 
possible picture, the thirty years of growth and change between 
1945 and 1975 must be studied. Five periods are discernible in 
the national and NIH biomedical research endeavors : 



5-1 



1945 - 1950 This was a period of retraining of physicians 
trained in war-time accelerated programs with 
little or no post M.D. training during the 
war years. Residency programs flourished after 
1945 and an intense desire for additional scien- 
tific knowledge was created. 

1950 - 1955 At the beginning of this period total annual 

federal biomedical investment was about seventy- 
five million dollars with industry supplying 
a like amount. An increased scientific base 
for medicine and an increase in the training of 
scientists occurred during this period. 

1955 - 1967 This period was marked by a rapid and pheno- 
menal growth in the biomedical science estab- 
lishment, both at the National Institutes of 
Health and in its partners within the academic 
community- -principally the academic medical 
centers. This was a period of diversification, 
increase in number of institutes and a marked 
increase in speciality training in the clinical 
sciences. This enabled clinical medicine to 
take the fullest advantage of the discoveries 
which were being made in the basic science 
laboratories . 

1968 - 1971 A period of executive indecision about the 

funding of the biomedical research effort was 
followed by cutbacks in allocation of funds. 
Then a renewed commitment to biomedical re- 
search emerged. 

1971 - 1975 From the NIH viewpoint this was a period of 
leveling-off of some programs and of the in- 
creasing effect of inflation, consumerism, 
political indecision, the imposition of short- 
term goals. From the point of view of the 
academic institutions this was a period of in- 
decision, instability and uncertainty. 

In 1974 total health spending in the United States had risen 
to more than one hundred billion dollars. Of this amount more 
than four billion dollars was spent nationally in support of re- 
search programs in the biomedical sciences. Of this latter amount 
NIH provided approximately 1.7 billion dollars. This amounted to 
ten percent of the total federal research and development budget s 
but because of NIH's intense involvement in the academic communi- 
ty, more than forty-six percent of the academic community's 



5-2 



biomedical effort was funded from NIH alone. Thus the emergence 
of the dominant federal role in the affairs of the academic medi- 
cal centers has been well established with consequent political, 
social and economic implications. Shannon has pointed out that 
the dominant role of the National Institutes of Health in the aca- 
demic community is the cause of serious real as well as potential 
conflict. Within the present setting NIH cannot pursue its mis- 
sion for the support and development of biomedical research in an 
effective fashion without broad participation of academic science. 
Conversely, the integrity of the biomedical scientist's effort 
within academic institutions is largely dependent on healthy and 
vigorous programs at the NIH. But the objectives of the academic 
institutions are highly diverse, containing goals that relate not 
only to research but also to education and service. The objectives 
of the NIH are comparably diverse, but are characterized by goals 
that are definable in terms of disease. Thus the "potential for 
conflict and for negative consequences" is obvious (8). 

Before turning to a discussion of the quantitative changes 
revealed by the data collected in this task it is well to comment 
that qualitative changes have occurred which are impossible to 
measure quantitatively. For example, the largest changes have 
occurred in schools which, before the emergence of federal support 
of biomedical research, had virtually no scientific effort. In 
other schools with a scientific tradition, much faculty time was 
available but the equipment and the technical support to develop 
the scientific research effort were lacking. As a result of in- 
creased availability of federal biomedical research funds many 
academic institutions have acquired the ability to do first-rate 
scientific research since World War II; however, this acquisition 

5-3 



does not show up in short-term data of the period 1964-1974. Sec- 
ond, institutions have become weaker in their ability to control 
their own destinies. These institutions have been beset by num- 
erous economic, social and political problems which are only par- 
tially reflected by the quantitative data which we have presented. 
The federal shift to short-term goals and limited objectives, the 
trend away from investigator-initiated research to collaborative, 
multi -investigator projects only partially describes the pronounced 
changes occurring in the academic medical centers and universi- 
ties which are the site of most of the bioscience research endeav- 
or. Numbers will not depict many of these qualitative changes. 

Causation cannot be inferred from the trend data and the 
simple correlation techniques using data such as are presented 
here. More sophisticated techniques will only show relationships 
which must be investigated further in the institutions themselves. 
And none of these techniques can document the uncertainty and in- 
stability which characterizes the academic medical centers of 
1976. 

A note of caution should be sounded about the comparability 
of data obtained from the NIH and that from the questionnaire of 
the Liaison Committee on Medical Education (LCME) . These data 
are not strictly comparable for a number of reasons. Table 5-1 
was prepared to reconcile data from the two sources for the three 
most recent years for which data are available. Note that there 
is a one-year offset between NIH obligations and LCME reports or 
revenues to medical centers. Figures in the boxes are not strict- 
ly comparable due to the differences in accounting, largely due 
to differences in categorization. For example, before 1974, aca- 
demic medical centers reported research revenues from all federal 

5-4 



TABLE 5-1 

RECONCILIATION OF NIH BIOMEDICAL RESEARCH 
SUPPORT DATA WITH AAMC DATA 



NIH Obligations to All 
Performers 5 - 

Total of All Programs 

Total for Research Programs 
Awarded by Institutes and 
Research Divisions (I/RD) 

Total for Extramural 
Research and Contracts 
(I/RD) C 



Fiscal Year 
1971 1972 1973 

$l,212 b $1,506 $1,484 



974 



822 



1,190 



1,022 



1,242 



1,068 



NIH Obligations to Domestic 
Medical Schoolst for - 

Research Grants 

Research Contracts 

Total Research Grants 
and Contracts 

Federal Research Revenues Re 
ported by Academic Medical 
Centers^ 

Direct Costs 

Indirect Costs 

Total Federal Research 



367 

N.R. 



367 



1972 
$441 b 
117 



558 



438 
N.R. 



438 



1973 

$432 

120 



552 



448 
76 



524 



1974 

$493 

141 



634 



a. 



includes direct and indirect costs 
'millions of dollars. 



Source: NIH Data Book. 



includes regular grants, general research support grants, 
centers and resources and R § D contracts. 

may include general research support grants and awards from 
federal sources other than NIH. Source: LCME Questionnaire, 
N.R. = not reported by NIH. f includes schools of osteopathy. 



5-5 



sources (e.g., NIH, other DHEW, AEC , NSF , DOD) in a single cate- 
gory. On the other hand, NIH prior to 1972 did not report research 
and development contracts. With these and the footnoted caveats, 
it is apparent that LCME reports of research revenues are greater 
than, but of the same order of magnitude as comparable NIH reports 
of research obligation to medical schools. 



5-6 



A. SUMMARY OF FINDINGS 

From the viewpoint of the academic medical centers a number 
of trends were observed in the data collected. The major find- 
ings are as follows : 

1 . Trends in Total Operating Budget and Federal Research 
Funding . Between 1964 and 1974 the aggregate total operating 
budget of all academic medical centers increased 208 percent in 
current dollars and 82 percent in 1964 constant dollars. Federal 
research funding, on the other hand, increased 76 percent in cur- 
rent dollars but only 6 percent in 1964 constant dollars. (4A-1F, 
1G) . 

In 1965 federal funds for the direct costs of research ac- 
counted for 36 percent of the total operating budget of 86 estab- 
lished academic medical centers. By 1974 this had fallen to 23 
percent (4A-1G). The percent decline was more marked for centers 
less involved in research and for those in the northeast and south 
(4A-8A, 9A, 10A) . The percent decline in federal research funding 
was less marked for research-intensive and western centers (4A-5A 
and 12A) . 

Other Budget Expenditures . At the same time that federal 
biomedical research support as a fraction of total budget reve- 
nue was declining, academic centers received greater revenues from 



^References refer to appropriate tables or figures. See Results, 
Section. 



5-7 



state and local governments and from clinical and professional ac- 
tivities. State and local government funds as a percentage of 
total operating budget increased from 13 percent in 1965 to 20 
percent in 1974 (4A-1F, 1G) . Service Programs and Professional 
Fees increased from 4 percent to 23 percent over the same period. 
Because the completeness of professional fee reporting improved 
throughout the period of study, early data are inaccurate and 
there is understatement in each year. Nevertheless the clinical 
service function of academic medical centers has clearly increased. 
Larger amounts of clinical income are being earned and depended 
upon by the institutions as a source of revenue. By 1974 each of 
the two categories (state funds and professional fees) accounted 
for approximately the same dollar volume of revenue to the medical 
centers as did federal funds for direct costs of research. 

Indirect Costs . Indirect cost recoveries were closely re- 
lated to total operating budgets (up 212% in current dollars). 
Due to inflation indirect costs rose more rapidly than direct re- 
search costs so that by 1974 indirect costs equalled about 1/3 of 
the direct costs of federally sponsored research. 

Newly Established Medical Centers . Contrary to some expec- 
tations new centers have approximately the same general pattern 
of expenditures as well established centers, especially with re- 
spect to federal research funds. Total operating budgets of these 
new centers were, of course, lower than those of established cen- 
ters (4A-13A, 13B). 

2. Trends in Use of Funding Instruments by NIH/NIMH . Over 
the ten-year period under study the instruments used for the trans- 
fer of federal research funds tended to shift from research grants 
initiated by a single investigator to large multi-investigator 



grants (program projects and centers) and to contracts (4A-1D) . 
Research grants, in absolute terms of 1964 constant dollars, rose 
20% over the decade but declined slightly in terms of the percen- 
tage of total funds awarded (4A-1E) . Training grants, fellow- 
ships and faculty awards, on the other hand declined both propor- 
tionally and in absolute terms (191 less in constant dollars). 
These figures, of course, do not reflect the uncertain future of 
training programs, the mainstay of future biomedical research 
productivity. 

3. Changes in Faculty and Students . The data indicate that 
the academic medical centers have increased greatly in numbers of 
faculty and in numbers of all kinds of students - housestaff, 
graduate and other health science, as well as medical students - 
taught by these faculty (4A-1I). Medical students increased by 
17,500 (541) while all other categories of students increased 
10,800 (30%). There were 16,700 additional full-time faculty, an 
increase of 105%. More clinical faculty were added than basic 
science faculty due to increased medical center involvement in 
clinical programs. 

In the past ten years 30 new schools were established, and 
the remaining 86 schools showed a 47% increase in size and func- 
tion. Overall, there was a doubling of graduate student enroll- 
ment and a marked increase in other health science student in- 
struction. 

4 . Construction of Facilities . The federal government made 
grants for construction at an annual rate of about 50 million con- 
stant dollars (4B-1) . These grants accounted for approximately 
forty percent of each construction project to which federal money 
was contributed (although by law, up to 80% federal funds may be 



5-9 



used) . Additional construction - about 10 to 17 percent of the 
annual totals - was completed without federal assistance. 

5. Federal Research Funding and Medical Curriculum . An ex- 
amination of the possible impact of biomedical research funding 
on curriculum change indicated that, except for the initiation of 
a larger number of medical scientist training programs (concur- 
rent award of M.D./Ph.D. degrees) in research intensive and pri- 
vate institutions, there was no demonstrable effect of biomedical 
research funding on curriculum development. Since 1965, 65 of 84 
schools have undergone extensive curriculum changes mostly in the 
direction of interdisciplinary, system-oriented teaching which may 
be the result (although our data cannot prove this assumption) of 
an increased bioscience knowledge base. 

There was no demonstrable effect of biomedical research in- 
volvement on the initiation of programs for family medicine (4B-3) , 
That is, "research intensive" institutions were just as likely to 
initiate family medicine programs as those schools in which less 
research is done. Family medicine programs were more common at 
public schools. Accelerated curricula (M.D. programs less than 4 
years in duration) were more likely to be present at schools which 
had obtained federal special project grant support. 

6 . Federal Research Funding and Organization of Academic 
Medical Centers . Medical centers were found to be extremely var- 
iable in their organizational structure, in numbers of health 
science schools and affiliated hospitals comprising the centers, 
and especially in numbers and kinds of centers for scientific re- 
search and development. In spite of the fact that from 21 to 361 
of average center expenditures are in support of federal biomedi- 
cal research activities, we were able to find no relationship 



5-10 



between the complexity of academic centers and increased biomedi- 
cal research funding except in the top twenty-five percent of re- 
search intensive academic centers. These centers had established 
more research and development centers in response to the availa- 
bility of biomedical research funds. Further, those centers which 
had the greatest research involvement had as many hospital affilia- 
tions as those less involved in research (4B-2). 



5-11 



B. CONCLUSIONS 

1. Major Trends in Academic Medical Centers . Since World 
War II, the medical schools have evolved into large, complex aca- 
demic medical centers. During the 1950's and early 1960's, their 
major growth was brought about by the expansion of biomedical re- 
search performed by the medical faculties. This expansion occur- 
red in response to Congressional actions which recognized that the 
advancement of knowledge was a public purpose for which substan- 
tially increased levels of federal support were provided. For 
example, NIH obligations to medical schools for biomedical research 
rose from $8.2 million in FY 1950 to $618 million in FY 1973. There 
was a concomitant increase in graduate research training to pro- 
vide basic and clinical scientists to man the expanded research 
effort, with federal funds for this research training rising from 
$4.0 million in 1950 to $170.7 million in 1970. The proportion of 
medical school budgets provided by federal support for research 
and research training reached 45 percent in 1966 (Table 4B-6). Al- 
though funds for research and research training continued to in- 
crease slowly after 1966, they accounted for a smaller proportion 
of medical school budgets because of the increase in expenditures 
related to expanded teaching and patient care functions of the in- 
stitutions . 

In 1968, medical schools began a substantial increase in class 
size in response to the call for an increased number of physicians 
to meet greater demands for medical services. In addition, medical 



5-12 



school faculties took on larger responsibilities for the education 
and training of other health professionals and health workers re- 
quired to provide the increased level and complexity of medical 
services. The development of new schools (30 since 1964) and the 
expansion of existing schools was assisted by federal funds pro- 
vided for construction, student loans and scholarships, and ope- 
rating costs. Federal obligations for medical education purposes 
grew from $6.6 million in FY 1965 with the initiation of a student 
loan program to $199.4 million in FY 1974 (Table 4A-1F). Institu- 
tions also received expanded support for their educational programs 
from state funds, tuition, gifts and grants from the private sec- 
tor and endowment income. 

The demands for more medical services grew out of the reduc- 
tion in financial barriers by Medicare, Medicaid and a greater 
coverage of costs through other third-party payers. The success 
of biomedical research in making the prevention, diagnosis and 
treatment of disease more effective also stimulated a demand for 
medical care. The academic medical centers responded to this de- 
mand for more medical care by increasing their provision of medical 
services . 

The academic medical centers provide the natural locus for 
translation of new knowledge gained through research to patient 
care and between 1964 and 1974 they were in the final phase of 
their evolution from undergraduate medical schools to sophisticated 
providers of research, health education and advanced medical care. 
This transformation to patient care centers also modified the in- 
come sources of the institutions. During the ten years under study 
the proportion of operating budget derived from professional fees 
and service programs increased from 4% to 231 (Table 4A-1G). 

5-13 



More students and a major commitment to provide health ser- 
vice to their communities have changed the character and the mis- 
sion of the academic medical centers. The research emphasis has 
been diluted, but dilution has occurred because of other demands 
to which the institutions have had to respond and not because of 
diminution in research interest within the institutions. The 
critical question that is now posed is whether the expansion of 
student enrollment and the expansion of service without a propor- 
tional expansion of research in the nation's academic medical com- 
munity will, in the future, endanger the goal of improving health 
through the advancement of knowledge and the translation of know- 
ledge into medical service. 

2 . Current Fiscal Trends in Public and Private Academic 
Medical Centers . A study of "research universities" by the Ameri- 
can Council on Education has shown that those institutions, parti- 
cularly private ones, have been losing ground financially in re- 
cent years (9) . Data of the present study indicate that academic 
medical centers in general have not been so adversely affected 
(Figure 5-2). However, the rate of constant dollar increases in 
total operating budgets in the past two years has slowed. This 
slowing together with recession, inflation and other factors sug- 
gests that the academic medical centers may, like the "research 
universities", face a deteriorating financial condition in the near 
future. Comparison of trends of total budgets for 86 established, 
43 public and 43 private academic medical centers (Figure 5-2) shows 
that total budgets in private centers declined since 1973 while 
public centers have continued to increase revenues. This suggests 
that in fiscally stringent circumstances public and private cen- 
ters may respond differently. When pressed to support programs, 

5-14 



1 


1 

/ 


' 


M 










n 






X 
X 

x* 

X 




u 




LU 


X 

1 








1 

an 

Q- 








S 1 

OO ■ UjE 

UJ I 1— 1 
UJ I <=C» 


■3" 


UJ ~ 






u_ i i— ; 




=3 


X 
X 


-u V 

nr \ 


5 | J 




i 


X 
X 


UJ 


PROFE 


1 


o 




UJ 

cc: 
1 



savnoa inwisnod wei do snoitiiw 



. , 


Q 




LU 


1 t- 1 - 1 


1— 




U 




LU 


UJ *-> 


_l 




LU 




tfi 


CD LU 




^ 1=1 


CSL 


u_ «=c 


O 






<£ 







Ln 




CD 














i—\ 






LTv 

1 








i :j 




\ 






\ 


1 










w 






6 












■^ 






\ 












■<*. 






\ 


1 














/ 


















/ 














\ 




\ 

% 
t 




\ 






CO 
Q- 








% 




\ 




oo 

LU 
LU 

U 


3 




»* 
^ 




\ 

% 


\ 

















i 


L 








1 — 


Xfc, 




1 


i 


L 


_J 






UJ 








\\ 


<=c 






CD 






• 




\\ 








a 






\\ 






CD 






L_D 








1 








CO 




C_> 


\ * 




I 


CO 






_, 


'"-. 




15 


UJ 


CO 
LU 






«=c 


V 


UJ 


1— 


LL_ 








V 


oo 


h 


<C 


! 


CD 






o 


V 


LU 




Cd 






1 — 


Cd 


i ■ 


CO 


Q_ 








fci 




■ ■ 








1 




1 


1 











savnoa invisnod W6T do snoitiiw 




5-15 



public academic medical centers are better able to call on state 
revenues which furnish up to 321 of total budgets. This is a re- 
source not available to private academic medical centers to a sig- 
nificant degree (4% of average total budget in fiscal 1974) . Public 
medical centers also have received more favorable treatment than 
their parent universities at the hands of state legislatures until 
recently. However, there are many indications that fiscally hard- 
pressed state legislatures will soon have to adopt more stringent 
fiscal policies for all of their programs including the support of 
academic medical centers. By contrast, private medical centers 
do not enjoy a significant "buffer" of state funds and thus, like 
private research universities, may feel the financial pinch earlier. 
If this occurs it will become even more difficult, if not impossible, 
for academic medical centers to continue to share the costs of bio- 
medical research no matter how academically attractive research 
may be . 

3. Sources of Support for Biomedical Research . The data of 
this study and analysis by the Rand Corporation (10) indicate that 
federal biomedical research funds are appropriately and effectively 
used to conduct research. There is no evidence that research funds 
pay for teaching or other non-research activities. Actually our 
study indicates that biomedical research is partly subsidized by 
non-research funds from state governments, endowments, etc. This 
subsidization takes several forms. The first form is the direct 
institutional subsidy of research through the salary of the faculty 
member who conducts the research. The proportion commensurate to 
his effort on the research project often is not covered by the 
grant and must then come from institutional funds. As the Rand 
Study confirms, private institutions have traditionally charged a 

5-16 



more realistic share of the professor's salary to research grants, 
and public institutions are now beginning to follow suit where loc- 
al tenure rules and other circumstances permit. 

Second, academic institutions must maintain expensive facili- 
ties needed for research and continue to pay for these even when 
research grants are terminated. Third, full recovery of indirect 
costs of research are not possible under present rules (11) and 
thus are another form of research subsidy by academic institutions. 
Also, these indirect costs have increased more rapidly than direct 
research costs because fuel, administrative salaries and new re- 
quirements for affirmative action, animal care, more detailed ac- 
counting and reporting, etc. have been added to medical center ad- 
ministrative costs but not to research directly. 

The fiscal instability being induced in academic medical cen- 
ters because of their subsidization of biomedical research should 
be corrected. The proportional increases in the national invest- 
ment in biomedical research needed to diminish this steady drain 
on institutional resources is small. The advantage gained by 
stabilizing the institutions would be great. 

4. Role of Research in Medical Education . An active research 
effort has long been recognized as a necessary underpinning for 
the process of higher education in the United States. The decision 
that "universities" would be centers of continuing research in the 
U.S. was made a hundred years ago as Wolfle has emphasized (12). 
Medical education at that time was chiefly in proprietary schools 
and became solidly university-based only after the Flexner report 
was adopted (13) . Medical education thus became more allied with 
scientific research as it became centered in universities. This 
research emphasis was given impetus by the second World War and 



5-17 



the post-war expansion of research funding by the National Insti- 
tutes of Health. As Shannon has emphasized (8) the growth of in- 
stitutional research led to an expanded scientific knowledge base 
which was then translated into patient care through the undergrad- 
uate and post-graduate medical education process. Thus research, 
education and service have become inextricably intertwined with 
each dependent on the other for the ultimate success of the joint 
mission. Research, in this context, is essential to advance the 
cutting edge of science, to maintain quality in faculty and stu- 
dents and to provide the basis for advances in medical care. Be- 
cause research and education are so interwoven, allegations that 
research supports education are natural and recurrent. 

Separating these joint products is neither necessary nor ap- 
propriate. Biomedical research is intended to advance knowledge 
and to improve medical care. When biomedical research is conduc- 
ted in the academic medical centers it also contributes to the ed- 
ucation of physicians and other health professionals and the im- 
provement of medical care is advanced simultaneously with the ad- 
vancement of knowledge. If research were isolated from education 
achieving the goals of both enterprises would be made difficult. 
It is fortunate that the nation's investment in biomedical re- 
search has been able to yield a secondary benefit and through edu- 
cation to expedite the application of new knowledge to medical 
practice. However, the data presented in this report indicate that 
the nation's biomedical research investment has not covered the 
costs of supporting either the research or the educational revo- 
lutions which it has engendered. Instead, the institutions appear 
to have partially subsidized research while also supplying the re- 
sources necessary to provide the more sophisticated instruction 

5-18 



which the biomedical knowledge explosion requires. 

How much research is necessary for medical education is the 
source of some disagreement (14, 15) but both the Institute of 
Medicine and AAMC agree that some is essential. AAMC estimated 
that 37.3% of total educational costs were attributable to re- 
search activities required to maintain an adequate faculty. 

5 . Changes in Federal Mechanisms of Support to Academic Med - 
ical Centers . There has always been an emphasis toward targeted, 
disease-oriented investigation in federally supported biomedical 
research. Political and social pressures to increase this empha- 
sis have grown considerably, and in response, grants and particu- 
larly contracts have been directed toward satisfying these pres- 
sures. This response has been coupled with a move toward support 
of multi- investigator program projects and centers with the re- 
search directions more explicit than in the case of support for 
investigator-initiated research. Such directed research has un- 
doubted federal programmatic and managerial appeal. Institutions 
and faculty, however, develop their programs of research depending 
on their interests and capabilities. They make long-term commit- 
ments based on those interests. Many institutions, particularly 
the smaller ones, are increasingly unwilling to unbalance their 
own research, educational and clinical programs to respond to fed- 
eral initiatives. At the same time, their faculties wish to com- 
pete for grants and contracts so as to remain in the forefront 
of scientific knowledge development. This creates conflicts in 
academic institutions. A result of such pressures is seen in the 
response of research intensive medical centers to the availability 
of funds for research and development centers (Table 4B-2). Re- 
search intensive and private institutions have developed more 



5-19 



research and development centers than other medical centers. 

Further, institutions in responding to short-term initiatives 
and categorical disease programs must make long-term commitments 
for faculty. As initiatives change institutions cannot change as 
rapidly. To maintain the vigor of biomedical sciences there must 
be balanced opportunities for investigator- initiated research as 
well as for directed research. To meet this goal the federal gov- 
ernment should continue to call on the judgment of the academic 
community for developing new initiatives which will not unbalance 
the academic institutions. 

6 . Federal Role in Increased Training of Health Manpower . 
During the past ten years the enrollment of health science stu- 
dents has increased 50%, faculties have doubled and 30 new medi- 
cal schools have been started. It might be assumed that such a 
change in manpower production came about as a result of federal 
initiatives acting upon institutions which were able to respond 
because of previous federal support to research and research train- 
ing. This assumption is only partially true. 

The 100% increase in faculty (Table 2A-1I) was made possible 
by the existence in training in 1965 of 5,500 graduate students, 
460 of whom became basic science faculty in new schools. Most of 
these basic science faculty members received research training at 
federal expense, but they also received state or local support 
(especially in the public schools) from stipends, assistantships , 
lower tuition and other awards. Basic science faculty, however, 
comprised less than a third of the incremental faculties, more 
than two-thirds of the new positions being clinical. Clinical 
faculty, as contrasted with basic science faculty, receive little 
or no federal support until late in their training. Thus, clini- 

5-20 



cal faculty receive two-thirds or more of their training at their 
own or state expense. 

As a result of all of these factors a large part of faculty 
training support was derived from state, private or personal sour- 
ces and the rest was federal support of teaching and training. 
Thus, the federal -local partnership, it must be emphasized, made 
possible the manpower training revolution. 

7 . Curriculum Change and Biomedical Research Funding . An 
examination of the possible impact of biomedical research funding 
on curriculum change indicated that "research intensive" institu- 
tions had a larger number of medical scientist training programs 
(concurrent award of M.D. and Ph.D. degrees). However, except for 
this finding, our data do not unequivocally demonstrate an effect 
of biomedical research funding on curriculum development. Since 
1965, 65 of 84 schools have undergone extensive curriculum changes 
mostly in the direction of inter-disciplinary, system-oriented 
teaching. The expansion of the body of biomedical scientific know- 
ledge which resulted from federal research programs of the 1950' s 
undoubtedly stimulated the curriculum change. The implementation 
of curriculum changes required the larger numbers of scientifically 
well-trained faculty who were fortunately available as a result of 
earlier expansion of training programs. A corollary of this fact 
is that the amount of biomedical research and research training 
affects the education of the faculty and markedly influences the 
type of educator trained. It is apparent that this influence has 
been beneficial both to science and to patient care. 

There is no demonstrable relation of biomedical research in- 
volvement to the initiation of programs for family medicine. That 
is, "research intensive" institutions were just as likely to start 

5-21 



programs in family medicine as those schools in which less re- 
search is done. The Rand Study (10) also emphasizes that there 
is no relationship between research funding and career choice of 
students. Finally, "research intensive" medical centers have re- 
sponded the same as centers less involved in research to pressures 
for service to patients and increased manpower training. Both 
types of medical centers have increased their use of community re- 
sources and affiliations with hospitals. 

8 . Effects of Research Funding on Medical Student Career 
Choice . A frequent impression voiced inside as well as outside 
the academic community is that research intensive institutions 
train too many researchers and not enough "real" doctors , do too 
much research, do not take care of the sick and so on. We have 
found just the opposite to be true. Research intensive medical 
centers are just as likely to train students for family medicine 
or general practice (Table 4B-3) , do not inhibit medical students 
in their choice of primary care careers (10) and have as many 
community hospital affiliations as those medical centers less in- 
volved in research (Table 4B-2). 

The data are not available with which to answer definitively 
the question whether research intensive medical schools emphasize 
the training of researchers rather than primary care physicians. 
Recent data from American Medical Association files approximate 
the answer although the AMA does not distinguish whether a re- 
search physician affiliated with a medical school also provides 
clinical care or not (16) . These data show that 87% of the 1965- 
1969 graduates of the top 21 schools in research involvement en- 
tered clinical practice while 10% listed research and teaching as 
their primary activities. In contrast 94% of graduates of schools 

5r22 



least involved in research entered clinical careers and 41 were 
primarily in research and teaching careers. 

There seems also to be a current notion that research train- 
ing can be safely turned off when fewer researchers are "needed" 
and on when more are wanted. The Rand Study (10) suggests such 
is not the case. Certainly clinical research training is at best 
a delicate institution demanding patients, ongoing research pro- 
grams and a continuous supply of trainees to participate in these 
programs. Once the program stops, the physicians pursue other 
options and the patients are dispersed, much time and money are 
needed to start the program again. Clinical research, contrary 
to popular opinion, is not something any physician can do. In 
fact, it is as complex, as demanding and much more regulated than 
is basic research. Clinical research is the translation of basic 
discoveries to the bedside. As such it needs special encourage- 
ment . 



5-23 



REFERENCES 



1. Liaison Committee on Medical Education, Annual Medical School 
Questionnaire, Association of American Medical Colleges, 
Washington, D.C., 1975. 

2. Information and Instruction Handbook No. VI -A for Automatic 
Data Processing; Definitions and Specifications, Data Pro- 
cessing Section, Statistics and Analysis Branch, Division of 
Research Grants, National Institutes of Health, Bethesda, Md. , 
September, 1974. 

3. Selected Information on Health Professions Schools, Part II: 
BHM Support by Program, FY 1965-74, Program Management Infor- 
mation Section, Bureau of Health Manpower, Health Resources 
Administration, Bethesda, Md., 1975. 

4. Nie, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner and D.H. 
Bent, Statistical Package for the Social Sciences, 2nd e.d., 
McGraw-Hill, New York, 1975. 

5. Jaffe, S.A. § S.M. Adelman, Development of a Price Deflator 
for Biomedical Research, Contract No. N01-OD-2160, June 28, 
1974. 

6. Halstead, D.K., Higher Education Prices and Price Indexes, 

DHEW Publication No. (OE) 75-17005, Washington, D.C., 1975. 

7. Basic Data Relating to the National Institutes of Health, 
1975, W.T. Carrigan, (ed.), U.S. Government Printing Office, 
Washington, D.C. 

8. Shannon, J. "Federal Support of Biomedical Sciences," J. Med . 
Educ . , in press . 

9. Dimensions and Trends in Health-Related Research Funding in 
Universities, (Task 1) Contract No. NOl-PP- 5- 2159 , American 
Council on Education, Washington, D.C, 1976. 

10. The Impact of Federal Funds Upon University Life Science De- 
partments and the Impact of Federal Health Related Research 
Expenditures upon Academic Medical Centers , (Task 3) Contract 
No. NOl-PP- 5- 2159, Rand Corporation, Santa Monica, California, 
1976. 

11. A Study of Indirect Costs of Research (Task 4), Contract No. 
NOl-PP-5-2159, American Council on Education, Washington, D.C, 
1976. 

12. Wolfle, D. , "The Home of Science," McGraw-Hill, New York, 1972. 



13. Flexner, A., "Medical Education in the United States and 
Canada", Updyke, Boston, 1910. 

14. Institute of Medicine, National Academy of Science: "Costs 
of Education in the Health Professions, January, 1974. 

15. Committee on Financing of Medical Education: "Undergraduate 
Medical Education: Elements, Objectives, Costs", Association 
of American Medical Colleges, October, 1973. 

16. Martin, B.C., Medical School Alumni, Aspen Systems Corporation, 
Rockville, Maryland, 1975. 



PART IV 



The Effect of Federal Biomedical Research Programs 
on Academic Medical Centers 



A. P. Williams, G.M. Carter, D.S.C. Chu , S.B. Coleman, 
A.P.Massell, C.R. Neu , R.L. Rasmussen, W.H. Rogers 



The Rand Corporation 



PREFACE 



This report was prepared for the President's Biomedical Research Panel under 
contract from the National Institutes of Health of the Department of Health, Educa- 
tion, and Welfare ( NO l-PP-5-2159). The contract was administered by the American 
Council on Education (ACE) and provided for research by a consortium of that 
organization, the Association of American Medical Colleges, and The Rand Corpora- 
tion. 

The purpose of the research was to assess the effects of federal biomedical and 
behavioral research programs on institutions of higher education. This report, 
which examines the effects of those programs on academic medical centers, is one 
of four basic components of the research effort. The other three reports are: 

Lyle H. Lanier and Ivars Zageris, A Study of Financial and Educational 
Trends in Research Universities, in Relation to Federal Funding of Health- 
Related Research — 1964-1974, American Council on Education, 1976. 

T. E. Morgan and D. D. Jones, Trends and Dimensions of Biomedical and 
Behavioral Research Funding in Academic Medical Centers: 1964-1974, As- 
sociation of American Medical Colleges, January 1976. 

David E. Drew and John G. Wirt, The Effects of Federal Funds upon Selected 
Health-Related Disciplines, The Rand Corporation, R-1944-PBRP, 1976. 

A fifth component of the research, prepared jointly by the research staff of ACE, 
AAMC, and Rand, summarizes the major findings of the four basic reports. This 
summary appears as an appendix to the report of the President's Biomedical Re- 
search Panel to the President and the Congress. The portions of this summary that 
relate to the Rand analysis of academic medical centers appear in the Summary to 
this report. Since the body of this report describes the methods of analysis in some 
detail, we recommend that the Summary be read prior to the major substantive 
sections (II through V). 

Although this study was performed for the President's Biomedical Research 
Panel, it draws heavily on data collected for an earlier study of academic medical 
centers for the Health Resources Administration and the Office of the Assistant 
Secretary for Planning and Evaluation, Department of Health, Education, and Wel- 
fare (NO l-MB-24196). It also uses material from a concurrent study for the Office 
of the Director, NIH (NO l-OD-5-2127). 

This report should be of interest to those concerned with federal biomedical 
research policy and higher education as well as to those involved in policy research 
and evaluation. 



iii 



SUMMARY 



This report examines the effects of evolving federal research policies and pro- 
grams on the nongovernment institutions we call academic medical centers. Aca- 
demic medical centers include, in addition to medical schools, at least one but 
usually several major teaching hospitals, and often one or more semi-autonomous 
research institutes. The implementation of federal biomedical research programs is 
affected by all organizational components of a center. 

Just as it is necessary to consider the larger organizational complex, the academ- 
ic medical center, in assessing research program effects, so it is necessary to take 
account of the instructional and patient-care functions of centers in examining their 
research activities. Very simply, this is because education, research, and patient 
care are conducted jointly within every center. Furthermore, some of the same 
resources are used in producing the three classes of outputs of education, research, 
and care. In some instances, all three are produced simultaneously, but in very few 
instances is the production of one irrelevant to the production of at least one of the 
others. Most important to the federal government and the institutions, the cost of 
producing any one class of outputs depends on the amounts of the others that are 
being produced. 

Naturally, the mix of educational, research, and patient-care functions that are 
conducted by a center depends in substantial part on the incentives provided by 
external funding sources. In many cases, the most important of these is the federal 
government, which has programs concerned with all three classes of functions and 
their products. 

The major problem of analysis in this report is to sort out the effects of federally 
supported biomedical research from other influences on academic medical centers. 
We use multivariate analysis to assess the simultaneous effects of multiple factors — 
including federal research programs — that affect academic medical centers. 

Conceptual models of medical center activities have been developed to consider 
the effects of NIH research programs, and several statistical techniques permit the 
effects of other factors to be considered simultaneously. However, the models are in 
every case oversimplifications of complex processes, and the data are often only 
proxies for what one would wish to measure. Notwithstanding these limitations, the 
results of the analyses in most cases are strong enough and plausible enough for one 
to be confident of the direction, though not precisely of the magnitude, of the effects 
observed. 

Many federal research and research-training programs were directed at the 
basic medical sciences. Using individual departments as the unit of analysis, we 
examined the effects of various federal programs on enrollment and doctorate pro- 
duction. Although the general trend is one of doubling of size, there are, as one might 
expect, significant differences among institutions and across disciplines. Taking 
these into account, the effects of federal programs are still quite strong. Among 
federal programs, research-training funds received by a department appear to have 
the strongest effects on its enrollment and Ph.D. production. The amount of a 
center's general research support grants (formula grants based on a school's overall 
research funding) also significantly affected enrollment, and this probably reflected 



VI 



the overall research intensity of a center. After these two federal funding effects 
were controlled for, research funds to the individual departments had only a very 
small positive effect on educational program size. 

Our analysis indicated that the growth of basic science education programs may 
be attributed mainly to federal training programs that included student stipends. 
This raised questions regarding the likely effects of cutbacks in such funding. 
Preliminary analysis of trends and plans of department chairmen at ten medical 
schools suggests that the size of graduate programs in the basic science departments 
will be significantly related to the availability of funds expressly dedicated for 
student support. There appear to be only limited opportunities to build graduate 
programs with "self-supported" students. Since federal training funds are an impor- 
tant source of student support, cutbacks are likely to lead to changes in enrollment 
in the short run. Departments with substantial research funding have some limited 
capacity to support graduate students as research assistants in laboratory work. 

The role models of research faculty are thought by some to influence M.D. 
graduates of research-intensive medical schools to choose nonprimary-care special- 
ties. An examination of the graduates of ten medical schools indicated that the 
specialty choices of medical school graduates are affected very little by the intensity 
of research in particular departments or in the medical school as a whole, and there 
is no evidence at all to suggest that research intensity discourages a school's gradu- 
ates from entering primary care specialties. However, it is possible to predict with 
limited accuracy the choices of careers among broad categories of practice (internal 
medicine, other primary care, surgical specialties, other nonprimary-care special- 
ties) using characteristics of individual graduates (e.g., sex, undergraduate grade- 
point average, class standing in medical school). 

A research-intensive medical school environment positively affects the likeli- 
hood of a graduate's entering a research or academic career. However, only a small 
proportion of graduates of even the most research-intensive schools enter academic 
or research careers. Analysis of data on individual graduates of the ten sample 
medical schools indicates that research funding of the school where M.D.s received 
their undergraduate medical education was positively related to these choices, but 
the funding effects were much less important than individual characteristics such 
as class standing. 

It has been suggested that the research intensity of a medical center affects the 
size of its graduate medical education programs in various ways, in particular that 
high research funding encourages medical centers to expand the number of interns 
and residents beyond levels appropriate to the availability of patients for teaching 
purposes. Analysis of cross-section data for internal medicine yields no evidence to 
support this suggestion. Instead, it appears that the numbers of faculty and patients 
largely determine the numbers of interns and residents in the majority of depart- 
ments of medicine. 

We examined data on individual department size to determine the separate 
effects of research and educational programs. At the level of the departments (e.g., 
biochemistry), differences in faculty size across schools are related to differences in 
NIH research funding in a statistically consistent manner. In contrast, only for 
certain kinds of departments do teaching responsibilities appear to explain faculty 
size. However, where differences in numbers of faculty are related to teaching loads, 
the magnitude of the effect is much larger than that for research funding. That is, 



vii 



although for all departments the relationship between research program size and 
faculty size follows a more consistent pattern than the relationship between educa- 
tion program size and faculty size, the total effect of education programs on depart- 
ment size is greater than that of research. 

Analysis of data on departments of medicine in all institutions does not indicate 
that research funding significantly affects the clinical-care activities of departments. 
However, the limitations of the data and the models used for the analysis may 
explain the absence of observed effects. 

The dependence on federal biomedical research funds for faculty salaries varies 
greatly across institutions. Not surprisingly, those centers that have been most 
consistently successful in competing for research funds have tended to rely more 
heavily on this source of support. 

The only consistent trend in sources of support for faculty salaries is the in- 
creased reliance on practice earnings for the compensation of clinical faculty. How- 
ever, part of this change may be more apparent than real. It may result from the 
reclassification of some faculty from part-time or volunteer status to full time and 
from increased institutional accountability for practice fees. The real increase in 
revenue from this source due to expansion in the patient-care functions of academic 
medical centers and public-health insurance programs for the aged and needy is 
probably a one-time phenomenon. It does not represent a readily expandable source 
to replace research funds currently used for faculty salaries. 

Analysis of sources of support for individual faculty salaries does not reveal any 
consistent pattern of vulnerability to cutbacks in research and research-training 
funds that applies to all institutions. Although the proportion of faculty salaries 
funded from these sources does vary somewhat by the department (e.g., internal 
medicine, biochemistry) and the academic rank (e.g., full professor, assistant profes- 
sor) of the individual faculty member, most of the variation is accounted for by other 
factors. This does not confirm but is consistent with the hypothesis that differences 
in individual faculty, including, among other things, involvement in research, ac- 
count for most of the differences in their dependence on soft funds for salary. 

By far the most important determinants of average departmental faculty salary 
levels appear to be the type of department (anatomy, medicine, surgery, etc.), the 
region of the country in which the school is located, and the relative cost of living 
in the surrounding area. Taking these into account, there is a significant negative 
relation between NIH funding and faculty salaries. In departments with high levels 
of NIH funding, salaries tend to be lower, with this effect more pronounced for junior 
faculty than for senior faculty. Salaries of department chairmen, however, are 
positively related to levels of NIH funding. 

When changes in faculty salaries are related to changes in NIH funding, the 
results are not as clear. In clinical departments there appears to be a positive 
relationship between increases in grants and increases in full professors' salaries. 
This relationship does not seem to hold, however, for associate or assistant profes- 
sors. Salary increases for assistant professors appear to be very vulnerable to de- 
creases in NIH training grants. In basic science departments, no interpretable 
pattern emerges. 

Federal research programs may vary in size over time and medical centers may 
be more or less successful in competing for them from one year to the next. We 
analyzed data on institutional budgets to determine the effects of such changes from 



vm 



year to year. Changes in NIH funding seem to have only mild effects on funding from 
other sources. Where consistent effects are observed, they seem to be associated with 
increases rather than decreases in NIH funding. This suggests that other funding 
sources are not generally available to compensate for shortfalls in NIH funding. 
Neither is there any evidence of significant "multiplier" effects; increases in NIH 
funding do not seem to attract substantial funds from other sources. In general, the 
models relating changes in various categories of funding for the institution to 
changes in another single category (the effects of revenue from tuition, patient fees, 
NIH support, etc. on revenue from foundations) are not of much predictive value. 
In most cases, they explain no more than 25 percent of the variance. 

Analysis of data on the allocation of institutionally controlled funds (tuition, 
capitation grants, etc.) does not indicate that such funds are treated as substitutes 
for funds generated by individual departments (e.g., research and training grants, 
patient fees). These results indicate that academic medical-center departments act 
as entrepreneurial units whose functions depend in substantial part on their ability 
to generate funds from outside sources. NIH and other public and private research 
funding agencies are important sources of department-generated funds, and practice 
earnings are of growing importance to all clinical departments (particularly those 
of the high-earning specialties). 

The central administration of a medical center appears to exercise only limited 
control over total department budgets, at least in the short term. There may be some 
asymmetry in the central administration's budget behavior with respect to increases 
and decreases in departmental research and research-training funds, but the 
asymmetry is different from what one might expect. That is, it appears that in most 
institutions in our sample, a department may obtain more institutional funds by 
increasing its research support. However, a department's loss of research funds does 
not appear to have a significant effect on its allocation of institutional funds. This 
indicates that individual departments may be quite vulnerable to research funding 
cutbacks because the institution has little flexibility to compensate departments for 
such losses. 

Another inference that may be drawn from our analysis relates to the question 
of whether research funds "subsidize" the education programs of medical centers. 
No evidence has been found that research funds supplant institutional funds that 
are generated by, or that would normally be used exclusively for, the training of 
undergraduate M.D.s. That is not to say that the character of M.D. education pro- 
grams is not influenced by the presence of a research effort. 

From the point of view of the federal government, an important attribute of the 
research being performed in an academic setting under NIH and ADAMHA sponsor- 
ship is the Institute, or program within the Institute, that sponsors the research. The 
sponsor describes, at least in part, the disease entity or normal process that will be 
better understood as a result of the research. Budget allocations among programs 
are made on the basis of the health problem being studied. The review processes for 
NIH and ADAMHA are administered separately. Differences among the federal 
programs may have different effects on the different parts of the university. 

Federal programs also affect the research activity of individual scientists. To 
assess the effects of federal programs on scientific activity, it is necessary to examine 
changes over time in that activity; and the sponsoring agency is not an adequate 
description of research for this purpose. There is not enough detail to detect many 



IX 



kinds of changes that may have occurred. In some areas of science, research that is 
very similar in its methods, knowledge base, and scientific goals is sponsored 
through programs of different Institutes. Changes over time in the sponsoring agen- 
cy do not always signal real changes in scientific activity because there have been 
changes in the perception of the areas of basic research that are relevant to certain 
diseases. In addition, it might be possible for an investigator to influence the assign- 
ment of his application to an Institute by adding a disease-oriented window dressing 
to his application, without changing the scientific content of his work. To avoid these 
problems, scientific activity has been classified solely on the basis of its scientific 
content as described in the NIH IMPAC file. 

Changes in the relative funding levels of the various federal programs will have 
a greater effect on the parts of the university that are most heavily involved in the 
programs being enlarged or cut back. The analysis shows significant differences in 
the academic settings of research among institutes of NIH and ADAMHA. 

Applications assigned to the National Institute of General Medical Sciences 
(NIGMS) are much more frequently from a university science department than from 
a basic science department of a medical school. Applications for the National Heart 
and Lung Institute (NHLI) and the National Institute for Arthritis, Metabolism, and 
Digestive Diseases (NIAMDD) are most frequently from clinical departments of 
medical schools and, in addition, applications from basic science departments are 
much more frequently from the medical school side of the university. The situation 
for basic science at the National Institute of Child Health and Human Development 
(NICHD) is reversed, probably because of population and psychology studies. How- 
ever, applications come to the National Cancer Institute (NCI) from each component 
of the university in the same proportion as total NIH applications. 

ADAMHA receives 75 percent of psychiatry department applications but only 
10 percent of applications from the rest of university and medical school depart- 
ments. Within ADAMHA, departments of the medical school other than psychiatry 
have a higher than expected proportion of their applications going to NIAAA and 
NIDA, while university departments are more likely to apply to NIMH. Appli- 
cations from departments of psychiatry and health professions schools go to each of 
the three Institutes of ADAMHA in the same proportion as the total number of 
applications. 

Applications to ADAMHA were disapproved at a much higher rate than appli- 
cations to NIH. ADAMHA disapproved 52 percent of new applications and 22 per- 
cent of renewal applications, while NIH disapproved 33 percent of new applications 
and 22 percent of renewal applications. The rate of approval of NIH applications 
increased slightly during the 1971-1975 time period, while no trend is observable for 
ADAMHA. 

Applications to NIH from basic science departments of medical schools and 
graduate departments have a higher approval rate than applications from the rest 
of the university. Applications to ADAMHA from the basic science departments of 
medical schools also have a higher approval rate than those from the rest of the 
university, although the university graduate departments do not. 

The percent of approved NIH applications that are funded is the same for each 
component of the university. Approved applications to ADAMHA from graduate 
departments have a slightly smaller chance of being funded than approved appli- 
cations from the rest of the university. 



The scientific classification system from the IMPAC file describes NIH proposals 
for research project grants along four axes: Discipline and Field, Body System, 
Research Materials, and whether or not the project is drug-related. This system 
permits construction of a typology of biomedical science by grouping applications 
that have similar descriptions in the scientific classification system into clusters that 
constitute subfields of biomedical research. Since each grant may receive several 
codes from each axis, the detailed descriptions of most projects are unique. Never- 
theless, some sequences of codes appear repeatedly on different applications and 
describe the research subfield to which the application can be assigned. 

A set of 50 clusters has been identified by a new clustering algorithm developed 
for this purpose. Most of these clusters may be easily classified as a subarea of one 
of the interdisciplinary cluster panels assembled by the President's Panel for Bi- 
omedical Research. 

An examination of the applications in these clusters shows that each component 
of the university performs a unique role in the spectrum of biomedical research. 
Clinical studies are performed almost solely by members of the clinical science 
departments of medical schools. The exception is clinical developmental studies that 
are performed in all components of the university except for the basic science 
departments of the medical school. The research performed in basic science depart- 
ments of medical schools differs significantly from research performed in graduate 
schools; the former is more likely to be drug-related, based on a body system, or 
related to a medical specialty. Some fields of biomedical research within chemistry 
are performed in university departments but hardly ever in medical schools. The 
medical school's clinical departments and the university graduate departments are 
involved in behavioral research, but this is almost totally absent from the basic 
science departments of medical schools. 

Federal program priorities affect the scientific activity being performed under 
research project grants by funding some fields at a higher rate than others. One way 
to examine this effect is to compare actual funding rates with what would have 
happened if the study-section priority scores were the only criteria used in awarding 
applications in each year. On the average, 16 percent of total decisions made be- 
tween 1971 and 1975 to fund or not to fund applications were different from what 
they would have been if scientific merit were the only criterion. 

Federal programs might have an additional effect on research by influencing the 
kinds of research that scientists propose in their applications — i.e., a scientist with 
an opportunity to work on several problems might write a proposal for the one that 
he believed had the highest chance of funding. Changes over time in the number of 
competing applications in each cluster is one way to describe the demand for re- 
search support in the scientific field. In addition to the perceived likelihood of 
obtaining research support, demand for research support in an area depends on the 
scientific opportunities available in that area and on the number of available scien- 
tists who possess the training and ability to work in that area. There is no way of 
directly measuring any of these quantities, so surrogates are used for each one. 

The proxy for funding levels due to governmental priorities is the difference 
between actual grants awarded in a cluster and the number of grants that would 
have been funded if the priority score had been the sole determinant of funding. The 
proxy for the scientific opportunities available in a field is developed from the 
distribution of priority scores in each cluster. For the supply of manpower, the data 



XI 



are the number of scientists working on grants in a cluster. The analysis provides 
some evidence of a response by the scientific community to availability of funding. 
For every application awarded beyond the nominal cutoff line in a research subfield 
in 1971 and 1972, two additional applications were received in 1974 or 1975. How- 
ever, the variable that appears to dominate in year-to-year changes is the availabil- 
ity of manpower. 



ACKNOWLEDGMENTS 



This study would not have been possible without the cooperation of the adminis- 
trators and faculty of ten academic medical centers that served as the sample 
institutions for this and our earlier research on federal program effects. They have 
in all instances been generous with their time, candid with their comments, patient 
in explaining the operations of their centers, and trusting in granting us access to 
sensitive data that were essential to thorough analysis. 

The active cooperation of the Association of American Medical Colleges (AAMC) 
has been vital to our research on the academic medical community over the past four 
years, of which this study is but one part. We are particularly grateful to John A. 
D. Cooper, President of the AAMC, for his continued interest in and constructive 
criticism of our research. Paul Jolly and Jesse Darnell were instrumental in helping 
us to understand and use data from the AAMC's Institutional Profile System, which 
was very valuable to this study. The American Medical Association provided data 
on the careers of medical school graduates from their Master File of Physicians. 
Solomon Eskenazi and William B. Casey of NIH and David F. Kefauver of ADAMHA 
were instrumental in providing and helping us to interpret data on extramural 
grants of those two organizations. 

Our close working relationship with Lyle H. Lanier of ACE and Thomas E. 
Morgan of AAMC, the project leaders for the ACE and AAMC components of the 
research for the Panel, was invaluable throughout the course of the study. We also 
benefited greatly from the assistance of the study advisory committee: Steven Mull- 
er (chairman), Alfred J. Bollett, Stuart Bondurant, John R. Hogeness, Doris Merritt, 
Leslie T. Webster, and John T. Wilson. 

Throughout the study, we have benefited from our interaction with Charles 
Lowe, Jarold Kieffer, and Charles McKay, the Executive Secretary, Staff Director, 
and our project monitor of the Panel. The comments and questions of Panel mem- 
bers during our briefings to them helped to sharpen the focus of our final report. 

Performing a study of this magnitude in a short time has required the collection, 
management, and processing of large quantities of data from diverse sources. In this 
area the efforts of Kent Brown, Henry Corona, Misako Fugisaki, Heather Hanunian, 
and Clara Lai were crucial to our analysis. Kenneth Barker, Bruce Bennett, Douglas 
Campbell, Dolph Hatch, John Lema, and Esther Uyehara provided valuable pro- 
gramming and other assistance in our research. Our final report has also benefited 
from the comments and criticism of Stephen Klein and Tom Lincoln who reviewed 
an earlier draft. 

The editing and coordination efforts of Helen Turin have been invaluable. Mari- 
lyn La Prell maintained an up-to-date computer-based version of our report through 
dozens of separate revisions by the various authors, and Beverly Westlund coor- 
dinated the final production process. 

While the success of this project owes much to those mentioned above, the 
authors bear full responsibility for any errors of commission and omission that 
remain in the analysis. 



xui 



CONTENTS 



PREFACE iii 

SUMMARY v 

ACKNOWLEDGMENTS xiii 

GLOSSARY xvii 

Section 

I. INTRODUCTION 1 

Data Sources 2 

Analysis 3 

Organization of the Report 5 

II. FEDERAL BIOMEDICAL RESEARCH FUNDING AND THE 
EDUCATIONAL PROGRAMS OF ACADEMIC MEDICAL 

CENTERS 7 

Basic Science Department Enrollment 7 

Data Sources 9 

Trends 10 

Further Analysis 10 

Summary of Findings 14 

Changes in Student Financial Support 14 

Conceptual Model and Policy Questions 14 

Data Sources 15 

Data Interpretation 15 

House Staff Size 17 

Implications of Patient Availability Constraints 18 

Data Sources 19 

Problems with Proxies 21 

Patterns and Trends 21 

Empirical Technique 25 

Analysis and Results 26 

Summary of Results 30 

Appendix A 30 

Appendix B 41 

III. THE EFFECTS OF FEDERAL BIOMEDICAL RESEARCH 
PROGRAMS ON THE CHARACTERISTICS OF 

DEPARTMENTS 43 

Determinants of Department Size 43 

The Model 43 

Data and Results 44 

Faculty Involvement in Clinical Care 46 

Objectives and Limitations 46 



xv 



XVI 



Data Sources and Variables 47 

Analysis and Results 47 

Limitations 49 

Reliance on NIH Funds for Faculty Salary Support 49 

Data Sources 50 

Analysis and Results 50 

Conclusions 55 

NIH Funding and Faculty Salary Levels 55 

The Data 56 

Analysis of Salary Levels 56 

Analysis of Changes in Salary Levels 60 

IV. THE EFFECTS OF FEDERAL BIOMEDICAL RESEARCH 

PROGRAMS ON INSTITUTIONAL FUNDING 63 

Alternative Sources of Revenue for Medical Schools 63 

The Model 63 

The Data 64 

Analysis 66 

Total Revenues from Student Tuition 74 

Conclusions 75 

Resource Allocation Decisions Within Academic Medical Centers . . 76 

The Model 76 

Analysis and Results 78 

Conclusions 81 

V. THE INFLUENCE OF FEDERAL BIOMEDICAL RESEARCH 

FUNDING ON RESEARCH AT UNIVERSITIES 82 

Federal Program Characteristics 83 

Assignment of Applications to Institutes 83 

Rate of Disapproval of Applications 83 

Rate of Funding of Applications 86 

Scientific Characteristics 87 

Data Base , 87 

Methodology 90 

Clusters of Applications by Scientific Field 92 

Trends in the Content of Biomedical Research 98 

National Health Priorities and the Content of Biomedical 

Research 100 

Research Content by Institutional Setting 104 

Conclusion 105 

VI. GENERAL CONCLUSIONS 107 



GLOSSARY 



ADAMHA ALCOHOL, DRUG ABUSE, AND MENTAL HEALTH AD- 

MINISTRATION 
NIAAA National Institute on Alcohol Abuse and Alcohol 

NIDA National Institute on Drug Abuse 

NIMH National Institute on Mental Health 

NIH NATIONAL INSTITUTES OF HEALTH 
NCI National Cancer Institute 

NEI National Eye Institute 

NHLI National Heart and Lung Institute 

NIA National Institute on Aging 

NIAID National Institute of Allergy and Infectious Diseases 

NIAMDD National Institute of Arthritis, Metabolism, and Digestive 

Diseases 
NICHD National Institute of Child Health and Human Development 

NIDR National Institute of Dental Research 

NIEHS National Institute of Environmental Health Sciences 

NIGMS National Institute of General Medical Sciences 

NINCDS National Institute of Neurological and Communicative Disor- 

ders and Stroke 



xvu 



I. INTRODUCTION 



The federal government relies heavily on nongovernmental institutions to carry 
out its policies in biomedical research. Although the first federal extramural re- 
search grants date back to 1918, the scope of present day involvement with universi- 
ties, medical schools, and teaching hospitals is the result of the rapid growth in 
federal biomedical research expenditures that began in the early 1950s. 

From this growth an interdependency has developed between the federal gov- 
ernment and the institutions that perform its research. On the one hand, the institu- 
tions have developed along lines that make them responsive to and reliant upon 
federal research funding. On the other hand, the federal government has designed 
programs on the assumption that these institutions will both perform the bulk of 
the research it seeks and train the scientists that federal programs will need in the 
future. As a result, the efficiency of federal government programs in biomedical 
research depends in substantial part on the efficiency of these nongovernment insti- 
tutions. 

The purpose of this study is to assess the effects of evolving federal research 
policies and programs on the nongovernment institutions that must carry them out. 
The particular focus of this report is on the institutions we call academic medical 
centers, each of which contains a medical school. The treatment of medical schools 
in a report separate from one dealing with "universities" is explained in part by the 
greater interdependency between federal biomedical research organizations and 
medical schools than between those organizations and universities as a whole. The 
separate treatment can also be attributed to the broader range of functions and 
greater organizational complexity of academic medical centers, of which medical 
schools are only one component. 

It is the presence of the joint functions of education, research, and patient care, 
rather than any particular formal organizational relationships, that determines the 
existence of an academic medical center. A medical school is the core component of 
every center, but the school often bears exclusive responsibility for only the under- 
graduate medical education programs. Sometimes several (but at least one) teaching 
hospitals share the responsibility with the M.D. degree-granting medical school for 
the clinical portion of medical education and have major responsibility for graduate 
medical education. The corporate relationships between medical schools and their 
teaching hospitals vary from outright ownership of the hospital by the medical 
school to a purely informal relationship between two totally separate corporate 
entities. The biomedical research functions of academic medical centers may be 
under the administrative purview of the medical school, its teaching hospitals, an 
affiliated but corporately distinct research organization, or some combination of the 
three. 

Whatever the organizational peculiarities of any academic medical center, the 
factors affecting the implementation of federal biomedical research projects are not 
confined to the medical school. Similarly, the effects of federal programs are not 
confined to any single organizational component of a center, no matter who bears 
the responsibility for administration of the federal research funds. Thus, it is appro- 



priate to focus an examination of the effects of evolving federal biomedical research 
policies and programs on the entire academic medical center. It is also necessary to 
take account of the education and patient care functions of centers in examining 
their research activities. Very simply, this is because education, research, and pa- 
tient care are produced jointly within every center. Some of the same resources are 
used in producing the three outputs of education, research, and patient care. In 
some instances, all three are produced simultaneously, but in very few instances is 
the production of one totally irrelevant to the production of at least one of the 
others. 1 Most important to the federal government and the institutions, the cost of 
producing any one group of outputs depends on the amounts of the others that 
are being produced. 2 

The mix of outputs of education, research, and patient care that are produced 
by a center depends heavily on the incentives provided by external funding sources. 
In many cases, the most important of these is the federal government, which has 
programs concerned with all three classes of outputs. 

The focus of this report on the effects of research programs does not imply any 
normative view about how much, where, or for what the federal government should 
spend research dollars. We are attempting to (1) provide an objective analytic de- 
scription of what has happened as a result of past federal research funding, and (2) 
present the analysis clearly enough that the reader may judge whether it is reason- 
able to extrapolate from the past effects to future ones. 



DATA SOURCES 

To analyze the effects of federal biomedical programs on academic medical 
centers, it is necessary to use data that range from strictly quantitative to highly 
subjective. The more subjective data (e.g., interviews with deans and department 
chairmen) are used to generate the hypotheses and develop the models of program 
effects, and the more quantitative data (e.g., federal expenditures, enrollment) are 
used in the actual hypothesis testing. 

Some important limitations on the analyses of the questions this report ad- 
dresses are imposed by the nature of the data that can be used. Although specific 
discussion of data constraints is best left for the sections that address separate 
analytic questions, the reader should have, at the outset, a general understanding 
of the data sources that were used in the overall report. 

The major source of subjective data on the operations of academic medical 
centers is an in-depth study often institutions done by The Rand Corporation from 
mid-1972 to mid-1974. That study was aimed at developing an understanding of the 
effects of a wide range of federal programs involving education, research, and pa- 
tient care on the internal operations and output of academic medical centers. 3 

1 An example of simultaneous joint production would be a case in which a patient is being treated 
with a new drug in a teaching hospital: The attending faculty physician is using the case to instruct house 
stafTand medical students in clinical medicine, he is providing therapeutic treatment to the patient (with 
the assistance of house staff), and he is collecting data for clinical research on the effects of the new drug. 

2 A more complete discussion of the implications of joint production is presented below. 

3 The final report of that study is presented in G. M. Carter, D. S. C. Chu, J. E. Koehler, and A. P. 
Williams, The Federal Government and Academic Medicine, R-1814-HEW, The Rand Corporation (forth- 
coming) This report contains a description of the procedure used to select these ten cases. 



Although this earlier study was concerned with many of the same questions as the 
present one, the focus was much broader. That is, federal research program effects 
were not the only matters of concern. To deal with the research programs more 
exclusively and to update our information on these institutions, we have revisited 
each in the course of this study. 

Detailed data on educational programs and graduates were obtained from the 
ten sample institutions. These include information on graduate programs in the 
basic sciences, including funding for students, and information on the background 
and medical school experience of the classes of 1955, 1960, 1965, 1969, and 1972, 
which we use (in conjunction with AMA data) to assess NIH research program 
effects on the careers of M.D. graduates. 

Data on the postmedical school education and career characteristics of graduates 
of the sample classes from the ten medical schools were obtained from the American 
Medical Association. These data are maintained from annual questionnaires sent to 
all M.D.s whether or not they are AMA members. 

Data on departmental budgets were obtained from the financial offices of the ten 
sample medical schools. These data were broken down by sources of support (e.g., 
NIH, NSF, foundation, hospital) and designated use of funds (e.g., research, graduate 
training, patient care). 

Data on research and research training grants from NIH and ADAMHA for all 
institutions were obtained from the IMPAC file maintained by the Division of Re- 
search Grants. Data for each academic medical center were compiled by aggregating 
grants to the medical school and its major teaching hospitals. 

Data on the patient care and graduate medical education activities of centers 
and selected clinical departments were obtained from the various issues of the 
Directory of Approved Internships and Residencies. For purposes of our analysis, we 
drew the boundary to include information on only "major teaching hospitals," as 
listed in the Directory. 

Data on educational programs of all centers were obtained from the various 
medical education supplements of the Journal of the American Medical Association. 
The data were originally extracted from the annual questionnaires of the Longitudi- 
nal Committee on Medical Education (LCME). 

Data on faculty size of medical school departments were obtained from the 
Institutional Profile System (IPS) maintained by the Association of American Medi- 
cal Colleges (AAMC). These data also came originally from the LCME questionnaire. 

Data on sources of funding for medical schools were also obtained from the IPS. 
To protect the security of these data, we performed the analysis using the computa- 
tional facilities of the AAMC. In all cases, we have taken care to protect the confiden- 
tiality of sensitive data on individuals and institutions. We report the analytic 
results without identifying the subjects, whether the population is the departments 
of an individual center, centers from our sample often, all centers on which we have 
data, or M.D. graduates of the ten sample centers. 



ANALYSIS 

All the questions addressed in this report have one thing in common: They call 
for sorting out the effects of federal research funding from other effects on the 



activities and outputs of academic medical centers. Conceptually, this requires a set 
of behavioral models of medical center processes and a large body of data on the 
funding, functions, and outputs of the centers. As a practical matter, the models 
used in the analysis are extremely simplifed representations of the complex pro- 
cesses that determine centers' behavior. In many cases, the available data are only 
crude proxies for the functions and outputs of interest. Even in the case of funding 
data, which are among the more precise, budgeting units and accounting conven- 
tions change over time so as to confound longitudinal comparisons. 

The analysis uses a number of multivariate statistical techniques that yield 
results understandable to laymen. In almost every case, the analytical approach to 
a question necessarily assumes a direction of causality by relating a group of inde- 
pendent variables, including NIH funding, to a dependent variable that is the par- 
ticular outcome (such as enrollment or faculty size) being examined. Such analysis 
cannot capture some complex interactions that occur in the real world. However, 
it does permit the assessment of NIH program effects while controlling for some 
other factors that are hypothesized to influence the outcome being examined. In 
some cases, such as the relations between enrollment and faculty size, we examine 
influence in both directions. 

In spite of the limitations of our models, we believe the results of the analysis 
provide appropriate bases for cautious inferences about the effects of federal bi- 
omedical research programs on academic medical centers. However, if the nature 
of the inference is to match the structure of the analysis, it is essential to distinguish 
between analyses that use longitudinal or time series data and those that use only 
cross-sectional data. 

Time series data conceptually provide the soundest basis for inferences about 
the effects of programs on particular institutions. These data permit one to analyze 
the relationship between, say, enrollment in a particular institution and that insti- 
tution's levels of NIH research and training grant funding over time. Since the time 
span of these data is usually limited, a number of institutions or subunits of an 
institution, such as academic departments, are analyzed together. Such analysis 
assumes that the pattern of relationships between the independent variables and 
the dependent variable is the same for all organizational units being analyzed, and 
remain the same over the time period being considered. 

Cross-sectional data analysis is appropriate for explaining sources of differences 
among institutions at a particular point in time — for example, why some institu- 
tions produce more Ph.D.s than others. It does not directly address questions about 
the effects of changes in a program on an institution, and, consequently, inferences 
about such effects should be made with considerable caution. Analysis of cross- 
sectional data, in contrast to that of time series data, assumes nothing about a 
constant pattern of relationships over time but instead assumes that the pattern of 
relationships is the same for all institutions being analyzed at the particular point 
in time that the data were collected. 

The explanatory power of the models will vary according to the structure of the 
model and the quality of the data used in the analysis, but it is important to 
recognize that overall prediction is not the most important criterion forjudging the 
quality of the models used. Instead, our major concern is to develop models that are 
likely to be most sensitive to the effects of federal biomedical research programs. It 
is simplest to explain the difference between these two criteria by example. If the 



objective is to maximize overall explanatory power of a model, one can introduce 
independent variables that capture the continuity of the system being observed — 
earlier year's enrollment to explain the current year's Ph.D. production or last 
year's faculty size to explain the current year's faculty size. However, if the objective 
is to make the analysis most sensitive to year-to-year changes in funding, a model 
that relates only differences in independent variables over time to differences in the 
dependent variable is more appropriate. The resulting difference equations would 
exclude from consideration all factors related to the continuity of the system of, for 
example, faculty salaries, and examine only the effects of, say, year-to-year changes 
in NIH funding and the cost of living. Although the overall explanatory power of 
a difference model is usually small because of the unaccounted-for short term pertur- 
bations, the results are sometimes less subject to misinterpretation. 

In most cases, we tried more than one analytic approach in addressing the 
questions in this report. In most cases, there is no simple correct choice of a best 
model because all require assumptions. The most obvious cases are the assumptions 
about stability of structure in time series models and the assumptions about the 
homogeneity of structure in cross-sectional models. Where we have restricted our 
analysis to the latter class of model, it was usually because of data constraints. In 
the report, we describe the results of all analysis other than the purely exploratory. 
However, in drawing our conclusions, we place most weight on the results of quan- 
titative analysis that we can easily interpret based on our overall understanding of 
academic medical centers. 



ORGANIZATION OF THE REPORT 

This report is organized into four substantive sections and a brief conclusion. 
Each of the substantive sections contains subsections that treat a category of ques- 
tions about the effects of federal biomedical research programs. In each subsection, 
which deals with a separate analytic task, we describe the conceptual model, the 
data used in the analysis, and the results. Then we offer conclusions regarding the 
federal funding effects on the particular activity or output being examined. 

Section II is an examination of the effects of federal research funding on the 
educational programs of centers. We consider basic science department graduate 
programs, student financial support, graduate medical education, and career deci- 
sions of M.D. graduates. In Section III we examine federal research program effects 
on the characteristics of medical school departments: the determinants of depart- 
ment size, factors that explain faculty involvement in patient care, the use of NIH 
research funds for faculty salary support, and the effects of research intensity on 
faculty salary levels In Section IV we look at federal research support in the context 
of the total funding for academic medical centers. We deal first with the apparent 
relationship between NIH funding and funding from other sources at the level of 
total center resources and then describe factors that influence allocation of institu- 
tional resources to individual departments. 

Section V examines the scientific content of proposals to NIH for research 
project grants from all components of the university, not just academic medical 
centers. The proposals are grouped into clusters representing subfields of biomedical 
research. We then examine the effects of the level of funding of a research area on 



subsequent applications for support in that area, and the parts of the university that 
perform different types of biomedical research. Section VI summarizes the findings 
of the other sections. Although we make no recommendations, this section contains 
a synthesis of our analyses that focuses on policy-relevant findings. 



II. FEDERAL BIOMEDICAL RESEARCH FUNDING AND 

THE EDUCATIONAL PROGRAMS OF ACADEMIC 

MEDICAL CENTERS 



The federal government is concerned about both positive and negative effects of 
its biomedical research funding on academic medical center educational programs. 
On the positive side, there should be adequate numbers and a satisfactory mix of 
biomedical researchers for the future. On the negative side, it is important to avoid 
possible adverse effects on educational programs outside the research sphere. 

Academic medical centers are a major source of graduate training for basic 
biomedical research. The medical school basic science departments train many of 
the Ph.D.s who will pursue biomedical research careers in academic and other 
settings. Of course, the vast majority of the country's research-oriented M.D.s re- 
ceive their undergraduate and graduate medical education in academic medical 
centers in the United States. Moreover, they almost always receive their research 
training as clinical fellows in major teaching hospitals or as postdoctoral fellows in 
medical school basic science departments. 

The mechanisms for federal influence on the centers' production of biomedical 
research manpower are its various grants for research and research training — the 
common research project grants, center grants, training grants, fellowships, general 
research support, and so on. Federal policy must be concerned with the response of 
centers to both increased and decreased funding in each of these areas. Apart from 
the direct effects of particular programs, the government needs to take account of 
the effect on research career education of overall changes in research funding to a 
center. 

Concern over adverse effects of research generally focuses on both undergradu- 
ate and graduate M.D. training. In the case of undergraduate medical education, 
research-intensive centers are thought to influence their graduates toward nonpri- 
mary care fields of the medical profession. In the case of graduate medical education, 
research is thought to increase the size of house staff programs above levels that can 
be justified by their contributions to patient care. 

In this section, we examine both the positive and the negative sides of the alleged 
effects of biomedical research funding. First, we analyze the effects of various classes 
of NIH funding on basic science department enrollments and Ph.D. production. 
Second, we examine effects of changing NIH policies toward training grants on the 
enrollment of and funding for graduate students in the basic sciences. Third, we 
examine the effects of research funding on the size of graduate medical education 
programs in medicine. Finally, we analyze data on physician specialties and career 
types to find evidence of the effects of the research intensity of the center in which 
they received their undergraduate medical education. 



BASIC SCIENCE DEPARTMENT ENROLLMENT 

Basic science departments of medical schools fill a multiple role. They contribute 



to medical education by teaching a portion of the undergraduate medical students' 
program, and they train masters and doctoral students in the sciences themselves. 
They undertake research on their own and in conjunction with private corporations. 
Basic science faculty members do not, as a rule, engage in direct patient care. - 

There are six main basic science departments: anatomy, biochemistry, microbi- 
ology, pathology, pharmacology, and physiology. In recent years, biophysics, cell 
biology, and genetics have also become important enough in some medical schools 
to warrant separate department status. The department of pathology is somewhat 
special because it is also a clinical department; that is, it is directly related to patient 
care. Pharmacology also has some clinical aspects. 

The Ph.D. graduates of basic science departments pursue a wide range of careers 
in industry, medical laboratories, higher education, and elsewhere. We have not 
attempted to examine the market for graduates in detail but we have found some 
evidence of an appropriate balance between supply and demand. The number of 
budgeted but unfilled full-time faculty positions in medical school basic science 
departments has stayed in the 3-6 percent range, an indication that departments are 
interested in hiring new faculty if they are good enough, but not a sign of critical 
demand. The number of postgraduate fellowships has been rising slowly, but in 
approximate proportion to the number of Ph.D.s produced. Field trips to several 
medical school basic science departments, undertaken as part of an earlier study, 
indicated that the departments usually placed their graduates after some effort. To 
a certain extent the department of pathology is an exception; there seems to be a 
continuing shortage of pathologists. 

Maintenance of the market for basic medical scientists constitutes an important 
motive for studying the role of federally funded biomedical research in the determi- 
nation of student enrollment. At the same time, it should be kept in mind that 
research funds influence other department characteristics, such as the size and 
quality of the faculty, the quality of the students, and the research itself. On the 
basis of historical records we have examined, it would be improper to draw conclu- 
sions about the desirability of increasing or decreasing research funding for the 
purpose of influencing enrollment. Also, it is difficult for us to say what might 
happen if the levels or distribution of funding were drastically changed. What we 
can do is provide some insight into how the existing institutions have been working. 

The basic instrument through which federal biomedical research funding affects 
enrollment is the department budget. As described in Carter et al., 1 budgetary 
inputs consist of department gifts and endowments, medical school deans' funds, 
private research grants and contracts, federal training grants, federal research 
grants and contracts, and university controlled funds. Competing claims on the 
budget include faculty salaries, administrative costs, research expenses, and student 
stipends. 

There are (at least) three theories on how basic science department enrollments 
are affected by available funds. The first is that departments base their admissions 
or cutbacks on whatever funding is supplied to them. Under this hypothesis, only 
funding in the current and immediate past year matters. The second hypothesis is 
that departments admit whom they please and provide support as best they can for 

1 Grace M. Carter, David S. C. Chu, John E. Koehler, Robert L. Slighton, and Albert P. Williams, 
Federal Manpower Legislation and the Academic Health Centers: An Interim Report, The Rand Corpora- 
tion, R-1814-HEW, April 1974. 



those who enroll. The third hypothesis is that admissions are determined by such 
a mechanism as the central administration or the anticipated market conditions. In 
all probability a combination of these factors is at work, with the situation differing 
from school to school. 

Site visits to ten medical schools, as reported in Carter et al., provide some clues 
tc what basic science department chairmen and medical school deans think hap- 
pens. That report says, 

The size of the basic science program is largely determined on a department 
by department basis, subject to little control by central medical school or 
university administration. The determination of size tends to follow the 
research strength of the departments closely since the reputation of the 
department determines the number of applicants who will apply and the 
availability of research grant funding for the support of graduate students. 
The number of places offered by departments is closely related in general to 
the student aid available. 

Data Sources 

Two main data sources were available for research. First, the Longitudinal 
Committee on Medical Education of the American Medical Association and the 
American Association of Medical Colleges conducts an annual survey of medical 
schools. Many of the results are published in the annual JAMA Supplement on 
Medical Education. From these we obtained for the years 1963-1974 (i.e., 1962-63 to 
1973-74) counts of enrollment in masters, doctoral, and postdoctoral training, and 
degrees granted in the important basic sciences departments (the six listed at the 
beginning of this section), as well as a total over all departments. This information 
was coded on a school-by-school basis. Although this survey is the best source of 
enrollment data that is readily available, it is not particularly well checked for 
accuracy and consistency from year to year. In the sample of ten schools that we 
surveyed in a previous Rand study, we found several discrepancies. However, the 
total number of errors seems to be small enough that the net effect on our results 
is negligible. 

The other main data source is the IMPAC file supplied by NIH. This contains 
all awards distributed in fiscal years 1967 through 1975. For the purposes of this 
research, we aggregated the awards by grant type within school, department, and 
year. We also recorded grants that were awarded to hospitals affiliated with each 
medical school department. 

For each of the departments, 16 types of funding totals were available to us. 
These were: 

(1) Research project funds (code R). These are usually small grants (less 
than $50,000) generally awarded to one faculty member (principal inves- 
tigator) for a specific research project. The funds usually pay a part of the 
faculty member's salary and may partially support some graduate students 
who serve as research assistants. 

(2) Training grants (code T). Awarded directly to the institution for the 
support of graduate students. 

(3) General Research Support grants (code S). Awarded directly to the 



10 



controlling institution according to a formula applied to all research grants 
obtained by that institution. These have been phased out since 1974. 

(4) Fellowship grants (code F). Awarded directly to students for their educa- 
tion. For the most part, these students have just earned their degrees and 
wish to pursue independent research. 

(5) Career development awards (code K). Awarded to junior faculty mem- 
bers who show unusual promise in research, giving them an individual 
means of support. 

(6) Program project (code P). Awarded to several faculty members for their 
research project. 

(7) Clinical research center (code M). Pays for bed costs and staff in a 
clinical research project. 

(8) Other. 

(9) to (16) correspond to (1) to (8) except that they apply to hospitals that 
are affiliated with the medical schools. In cases where a hospital was affi- 
liated with more than one school, funds were divided in the same ratio as 
the research funds that were awarded to the medical schools directly. The 
formula was applied within individual departments. 

We have combined R, K, P, and M grants into a single category, which we call 
research. We also made calculations separately with similar results. Since the two 
data sources overlap for the years 1967-74, our results apply to that period only. All 
funding levels are reported in units of a thousand 1964 constant dollars, with Hal- 
stead's Higher Education Price Index used as the conversion factor. 

Trends 

Table 1 shows the overall trends for enrollment and funding per 1974 medical 
school department. Only the six major departments are included. Enrollment has 
remained almost steady with some increase in microbiology, pathology, and phar- 
macology. Ph.D. production has increased steadily, reflecting the large increases in 
enrollment that took place immediately before 1967. In constant dollar terms, re- 
search funds have increased, but training grants and General Research Support 
grants have decreased, although some funds were restored in 1974. Hospital grants 
are small compared with other money. In most cases, they probably differ little from 
grants to clinical departments except that they generally go to faculty members who 
are primarily attached to the recipient hospital that has decided to maintain a 
research program administratively separate from the medical school. 

Further Analysis 

The enrollment data described above are listed by year, school, and department, 
which may be viewed as three factors that explain differences in enrollment, Ph.D. 
production, etc. The effects of these factors can be examined in an analysis-of- 
variance framework to determine their relative influence as well as interactions 



11 









Table 


1 












TRENDS IN BASIC SCIENCE DEPARTMENTS 
(Average Per Unit) 










Students 




$1000' 


's (1964 


dollars) 




Year 


Ph.D.s 


Enrollment 


Postgraduate 
Students 


Research Training 


Hospital 


GRS 


1967 


0.80 


9.7 


1.98 


124 


32.7 


6.06 


170 


1968 


1.23 


10.5 


1.99 


125 


30.8 


5.51 


189 


1969 


1 . 22 


10.3 


1.60 


122 


31.5 


5.14 


178 


1970 


1.31 


9.6 


2.24 


110 


26.7 


4.12 


158 


1971 


1.33 


10.7 


2.16 


114 


25.2 


3.64 


143 


1972 


1.41 


10.6 


2.94 


130 


25.3 


3.64 


131 


1973 


1.34 


11.1 


2.63 


127 


18.4 


3.71 


47 


1974 


1.45 


11.5 


2.62 


148 


27.1 


3.60 


174 



NOTE: "Units" are all departments at 1974 medical schools, so totals are 

(114 x 6) times values shown. The one exception is GRS units, which are 

schools. A given department is entitled to a small fraction of GRS funds 
only . 



among them on dependent variables of interest (e.g., enrollment, research funding). 
A school by department interaction, for example, means the extent to which depart- 
ments within schools differ in enrollments depending on the schools. Such a differ- 
ence might arise if the anatomy department was strongest (or largest) at one school 
and the biochemistry department was strongest at another. Table 2 shows an analy- 
sis of variance breakdown for enrollments, Ph.D. production, research, training, and 
GRS funds. 

The interpretation of Table 2 is that most of the differences observed are differ- 
ences between schools. Some schools are large and rich in every respect, others are 
small and poor. The dominance of this status quo prevents a decisive statistical 
investigation into the effect of funding on institutional quality. There are also large 
effects due to the tendency for some departments to be large in some schools but 
small in other schools. Frequently, the school by department interaction in enroll- 
ments is due to large (small) biochemistry and microbiology departments and small 
(large) anatomy, pathology, and pharmacology departments at the same school. 
However, this phenomenon did not carry over to research funding. 

We also noticed that time trends account for a substantial portion of the school 
by department by year interaction variance. Such trends describe the changes in the 
relative standing of departments within the school and the relative standing of the 
school within a basic science field. 

Another analysis we performed involves regressing enrollments (masters plus 
doctorate degree candidates) on the various funding types available in the same 
year. The departments were considered individually as well as collectively. Forty 
variables were available for the regression (30 in the case of departments considered 
collectively). The 4(J variables arise from taking five funding types (research, train- 
ing, fellowship, GRS, and other) by two routes (school and affiliated hospital) by four 
department types (the department itself, the rest of the basic science departments, 
the clinical science departments, and other departments). In the collective analysis 



12 



Table 2 
ANALYSIS UF VARIANCE RESULTS FOR BASIC SCIENCE DEPARTMENTS 1967-74 







df 






% Variance E 


%plained 






Term 


Enrollment 


Ph.Ds 


Res 


. Fund 


mg T 


raining Grants 


GRS 


School 




113 


44.3 


28.3 




50.5 




44.4 


67.0 


Department 




5 


9.9 


7.6 




2.6 




1.1 


-- 


Year 




7 


0.0 


0.0 




0.0 




0.0 


12.5 


Slope 




1 


0.0 


0.0 




0.0 




0.0 


3.6 


Residual 




6 


0.0 


0.0 




0.0 




0.0 


8.8 


School x Department 




565 


22.9 


20.2 




31.7 




34.1 


— 


School x Year 




791 


8.9 


11.2 




3.0 




4.4 


20.5 


Slope 




113 


2.5 


2.4 




1.6 




2.1 


9.3 


Residual? 




678 


6.4 


S.8 




1.4 




2.3 


11.2 


Department x Year 




35 


0.1 


0.2 




0.4 




0.2 


— 


Slope 




5 


. 


0.0 




0.0 




0.0 


-- 


Res iduals 




30 


0.1 


0.2 




0.4 




0.2 


— 


Sciiool x Department x Year 




13.8 


32.5 




11.8 




15.8 


— 


Slope 




565 


4.8 


5.8 




6.6 




7.0 


— 


Residuals 




3,390 


9.1 


26.7 




5.2 




8.8 


-_ 



there are only three department types, since basic science department funding 
covers both the department and its complement. 

It is not reasonable to include all 40 variables in the same equation because 
almost half the funding types are moderately and positively related to enrollment 
and most are not physically relevant. Nevertheless, one can draw some conclusions 
about the relative influences of the different funding types. The overall conclusion, 
based on regressing basic science enrollments as a total on the 30 relevant variables, 
is that the training grants to the basic sciences, research funds to the clinical 
sciences, and school funds are the most important quantities. Once these variables 
are controlled for, basic science research funding has little additional effect. 

Within the departments, we list the selected variables by decreasing significance 
(and order of entry). For anatomy, the significant variables were training grants and 
GRS grants. For biochemistry, GRS grants, training grants, and training grants in 
the other basic science departments were significant. For microbiology, research in 
the clinical sciences and training grants were important. Other limiting factors than 
money on enrollment in pathology reduce the explanatory power of the regression, 
but training grants are still significant. In pharmacology, training and GRS grants 
are the most significant, but research in the hospital and the school seems to play 
a part, as do training grants in hospitals of the clinical science departments. The last 
result may be a statistical aberration or it may describe a tendency for clinical 
hospital programs to draw resources away from pharmacology in a basic science 
setting. In the physiology departments, GRS grants, training grants, and research 
in the school and hospital physiology department were significant. The r 2 levels for 
these regressions range from 0.25 to 0.6. 

A similar set of regressions was run for Ph.D. production. However, the results 
generally parallel the enrollment results, so they are not discussed here. 

When we control for training grants and school funds, research funds have a 



13 



small but positive effect on enrollment in every basic science department studied 
and are statistically significant in two of the six departments. We conclude that 
although research funds may be important to the existence of a department that can 
train people, it is the training grants and school funds that actually determine how 
many are to be trained. Since this analysis establishes only a relationship and not 
a cause, we must caution that the effect could go in either or both ways: Training 
grants may cause enrollment or enrollment may cause grants, or institutional char- 
acteristics may give rise to both simultaneously. We explored the reasons for the 
regression results by regressing effects calculated in the analysis of variance tables 
on quantities obtained for similar units. For example, one of these regressions 
sought to explain the changes over time in enrollment at each of the schools as a 
function of school averages for enrollment, funding of various types, and funding 
change. One striking finding was that levels of enrollment are well explained by 
levels of funding, but changes over time are not. However, these changes are reason- 
ably well explained by basic levels of both funding and enrollment: The larger 
departments have added more students and Ph.D.s than the smaller departments. 
These conclusions apply to the individual departments and averages across depart- 
ments (within schools). Although this tendency may be an artifact of growth at equal 
rates across schools, it does indicate where most of the extra graduates are being 
trained. Since the technical theory supporting the regressing of analysis of variance 
effects is poorly developed, we caution against placing too much emphasis on these 
results; however, they suggest that the major features of enrollment differences 
across schools and departments are imbedded in long term characteristics of the 
institutions. 

We performed several regressions with 1973 enrollment as the dependent varia- 
ble, with previous years' funding levels and enrollment levels as predictors. Each 
department was considered separately. The results are similar for all departments. 
One or perhaps two of the previous years' enrollments explain the current year 
better than any of the research, training, or general research support award levels. 
The r 2 values for regressions on the last two years of enrollment alone range from 
0.47 (pharmacology) to 0.91 (biochemistry). With funding variables included in the 
regression, training grants contribute the most in additional explanatory power, 
although the contribution is only marginally significant and 1973 training grants 
are generally more important than those of previous years. Similar regressions 
using 1969 to 1972 enrollments as the dependent variables indicate somewhat 
stronger effects of research and training grants in the earlier years. The contribu- 
tion of funds to the change in department size is not statistically overwhelming. 
Because only about 100 schools are available for any given department and year, 
effects large enough to be policy relevant may not be statistically detected. 

For the earlier years we cannot tell whether the strong relationship between 
enrollments from year to year represents funding sources missing from these equa- 
tions or a tendency for departments to manage support for the students they want 
to have. We suspect that growth was erratic, depending on the availability of such 
scarce resources as staff students, and money. The year 1973 marks a turning point, 
since in that year schools were hard pressed to support the students they had. 
However, it appears they were able to find the money and have continued to expand, 
perhaps on a more conservative basis. New funding and enrollment figures should 
be consulted when available for an appreciation of more recent trends. 



14 



Summary of Findings 

Many federal research and research training programs were directed at the 
basic medical sciences. Using individual departments as the unit of analysis, we 
examined the effects of various federal programs on enrollment and doctorate pro- 
duction. Although the general trend is one of doubling of size, there are, as one might 
expect, significant differences among institutions and across disciplines. Taking 
these into account, the effects of federal programs are still quite strong. Among 
federal programs, research training funds received by a department appear to have 
the strongest effects on its enrollment and Ph.D. production. The amount of a 
center's general research support grants (formula grants based on a school's overall 
research funding) also significantly affected enrollment and probably reflected the 
overall research intensity of a center. After these two federal funding effects were 
controlled for, research funds to the individual departments had only a very small 
positive effect on educational program size. 

We also found that the graduate enrollments and Ph.D. production of the cen- 
ters respond to the long term levels of federal support but do not appear to be very 
sensitive to short term fluctuations. 



CHANGES IN STUDENT FINANCIAL SUPPORT 

Conceptual Model and Policy Questions 

The National Institutes of Health have supported research training in the bi- 
omedical sciences since 1938 as part of their mandate to study the physical and 
mental diseases of man. Since World War II, much of this support has been in the 
form of training grants to institutions, paying both the salaries of the faculty and 
the stipends of the students. Training grants were awarded for three types of activ- 
ity: pre-Ph.D. training in the basic sciences (e.g., anatomy, physiology); postdoctoral 
training of Ph.D.s; and postdoctoral training of M.D.s, often combining both labora- 
tory and clinical study. 

As the general level of NIH activity reached a plateau in the late 1960s, so did 
appropriations for research training. Some members of the executive branch began 
to ask whether federal support is needed to assure an adequate supply of high 
quality manpower in these areas. Secretary of HEW Weinberger attempted to refor- 
mulate the nature of federal support for research training, but before his programs 
could get off the ground, Congress passed the National Research Act of 1974. The 
act repealed all existing authorities for NIH support of research training and creat- 
ed a new program of National Research Service Awards. These awards may be 
fellowships given directly to individuals by the federal government, or grants to 
institutions who in turn select the recipients. Congress and the executive branch are 
still engaged in dialogue about how this act should be implemented. It is clear, 
however, that as a result of these discussions there may be significant changes in 
assistance for training in the basic biomedical sciences. It is therefore of interest to 
examine how changes in this support might affect the number of Ph.D.s trained. 

To better understand what the effects of changes in federal support might be on 
Ph.D. training, it is useful first to ask why the basic science departments in academic 



15 



medical centers have Ph.D. programs. First, and perhaps most important, is that the 
training of Ph.D. students is viewed as one of the principal activities of such a 
department. Second, it is widely believed that it is necessary to have a strong Ph.D. 
program to attract good faculty. Third, Ph.D. students play an important role in the 
research program of the institution by serving as junior professionals in a larger 
research team. Finally, in some of the basic sciences — such as anatomy — the stu- 
dents assist in the teaching program by instructing M.D. students. 

The number of graduate student positions that a department offers depends on 
the age of the department, its perceptions about long run demands for trained people 
in the field, the size of its faculty, the size and nature of its research effort, the 
amount and nature of its M.D. teaching responsibilities, and the number and quality 
of applicants for graduate positions in that field of study. The number of applicants, 
in turn, is a function of job opportunities for graduates in the field, and of the amount 
of scholarship support available. A large part of that scholarship support currently 
comes from the federal government, and changes in such support will obviously have 
an effect on the number of students trained. Changes in federal funding of depart- 
mental expenses, through their influence on faculty size, will also affect the number 
of trainees. 

Data Sources 

We are currently gathering the data on enrollment and sources of student 
support with which to investigate how large the effects of any particular change in 
federal support will be. Until we have enough of these data, we do have an alterna- 
tive way of gaining some understanding of how changes in federal support for Ph.D. 
training in the biomedical sciences would affect enrollment in those programs. That 
alternative consists of information provided by a "natural experiment" that oc- 
curred in 1973, when the executive branch attempted to cut back on training grant 
funds. During the fall of 1973 we collected data from the departments of anatomy, 
biochemistry, microbiology, pathology, pharmacology, and physiology in the ten 
medical schools with which we were working closely. These data focused on graduate 
student enrollments. They were collected after the termination of the traditional 
training grant programs and the inception of the so-called Weinberger fellowships 
had been announced, but before the court cases on impounded funds had been 
decided, and before Congress had passed the National Research Act of 1974. Thus 
departmental decisions were taking place in an environment in which department 
chairmen thought that future federal support for graduate training might be sub- 
stantially reduced. Moreover, most departments could not use their FY 1973 train- 
ing grant funds to appoint new trainees at the first-year level. 

We asked the departments to provide us the actual numbers of Ph.D. students 
in FY 1971 through 1974, the likely number for first-year students for FY 1975 if 
training grants continued, and the likely number if training grants were terminat- 
ed. Responses are tabulated in Tables 3 and 4. The tabulations include both depart- 
ments with training grants and those without such federal support. 

Data Interpretation 

It might be argued that we should confine our attention to departments cur- 
rently receiving federal funds (28 of 54 departments when these data were collected). 



16 



Table 3 
ACTUAL PH.D. MATRICULATION 



Field 

Ana Corny 

Biochemistry 

Microbiology 

Pathology 

Pharmacology 

Physiology 

TOTAL 



Number of 
Departments 

Responding 



Number of First-Year Students 
FY71 FY73 FY7A 



7 


15 


17 


17 


9 


35 


42 


49 


6 


31 


43 


15 


4 


8 


17 


8 


5 


14 


19 


11 


6 


38 


27 


44 




141 


165 


144 



Table 4 
PROSPECTIVE PH.D. MATRICULATION 



Field 



Number of FY75 with FY75 without 

Departments FY74 Training Training 
Responding Actual Grants Grants 



Anatomv 


6 


17 


17 


14 


Biochemistry 


8 


47 


49 


33 


Microbiology 


5 


14 


36 


14 


Pathology 


4 


8 


20 


6 


Pharmacology 


5 


11 


21 


10 


TOTAL 




134 


191 


105 



However, there are two reasons for believing the effects of the change of policy will 
not be limited to the departments directly concerned. First, our interviews with the 
department chairmen indicated that the prospect of federal funding provides a 
significant incentive to start or improve a Ph.D. training program. Thus, the reduc- 
tion of the federal training grants would affect both departments currently receiving 
training money and those without such support. Second, each school has a limited 
amount of funds available from its own resources to support the Ph.D. program. 
Reductions of federal support for one department may mean that this "dean's 
money" must be spread over more departments. Thus, cutbacks in one area can 
affect other departments within the school. 

It is important to recall that students matriculating in FY 1974 (see Table 3) 
could be covered only if the program had made the commitment before January 



17 



1973. Thus the effects of withdrawing federal support are visible in FY 1974 behavior 
of basic science departments. 

For the departments as a group, the cutback reduced first year enrollment by 
13 percent (see Table 3), returning enrollment to the FY 1971 level. However, much 
of the reduction was concentrated in departments of microbiology, pathology, and 
pharmacology; enrollments in biochemistry and physiology actually increased. 

Assuming the restoration of training grant funds was not to occur, the depart- 
ment chairmen predicted a further 22 percent decline in matriculation in 1975 (see 
Table 4). In contrast to Table 3, declines now occur in biochemistry and physiology. 

If these forecasts can be believed, and if these departments are typical of depart- 
ments in all academic medical centers, withdrawal of federal training funds would 
lead to a short run reduction of approximately 30 percent of Ph.D. matriculation. 
This reduction would not be spread uniformly among the various basic sciences. This 
natural experiment suggests that the steepest declines would occur in microbiology, 
pathology, and pharmacology, and the departments of anatomy, biochemistry, and 
physiology would be less affected. 



HOUSE STAFF SIZE 

In contrast with education in the basic sciences, clinical education fundamental- 
ly involves patient care. Treating patients in affiliated hospitals provides the experi- 
ence necessary to a medical school's clinical training, while clinical faculty, house 
staff, and medical students together contribute substantially to serving patients' 
health care needs. 

With patient care rather than research occupying the core of clinical education, 
it is unclear whether the role of research funding in determining house staff size 
would be similar to its role in determining basic science enrollments. It might be 
argued that the constraints and responsibilities imposed by patient-care activities 
limit the effects of research funding. For example, even if increased research funding 
contributed to a desire to expand house staff size, expansion might be inhibited by 
a lack of adequate numbers (or kinds) of teaching patients in affiliated hospitals. 

Alternatively, it might be argued that despite the need for teaching patients, 
house staff size remains subject to discretion: When research funding is high, major 
teaching hospitals in particular may expand house staff, increasing staff-patient and 
perhaps even staff-faculty ratios but exposing house staff to broader experiences in 
clinical research. A concern reflected in this hypothesis is that high research fund- 
ing may cause house staff size to expand beyond levels appropriate to the availability 
of teaching patients. 

This study examines the hypothesis that higher levels of research funding are 
associated with larger house staffs than would otherwise be found given the avail- 
ability of teaching patients. Data available for this study permit us to examine 
relationships between house staff size and measures of both patient availability and 
NIH research funding for six yearly cross-section samples of departments of medi- 
cine; measures of faculty size can also be included as control variables for three of 
the years. 

Although the analysis does not provide a test of the hypothesis that house staff 
size is constrained by patient availability, the possibility that institutions are con- 



18 



strained has important implications for interpreting the results presented here. 
Therefore, we begin with a discussion of the possible role of patient availability 
constraints. Then we describe the data, the empirical methods, and the results of the 
analysis. 

Implications of Patient Availability Constraints 

A number of factors may be taken into account by a clinical department, its 
medical school, and its affiliated hospitals in reaching a decision concerning house 
staff size. House staff provide valuable services in patient care, in research, and in 
undergraduate clinical training; a large and high-ranking house staff training pro- 
gram may even help attract respected clinical faculty. At the same time, house staff 
must be given adequate experience in treating various health conditions, they must 
have adequate access to faculty, and they must be supported financially. Future 
expectations, historical background, and other characteristics specific to individual 
institutions may also affect house staff size. 

The view that many factors affect house staff size is not necessarily inconsistent 
with the argument that patient availability imposes a constraint on house staff size. 
As shown in a highly simplified fashion by the solid line in Fig. 1, there may be a 
maximum acceptable house staff size that is consistent with each level of patient 
availability, but below which house staff size might vary among institutions. For 
example, suppose H A , H B , and H c are alternative levels of house staff size that would 
be desired, depending on research funding, faculty size, or other factors. An institu- 
tion with patient availability P* could choose house staff size H A (at point A) or H B 
(at point B), but could not achieve level H c because point C lies above the constraint. 

Although institutions that are not constrained by patient availability could be 
very responsive to effects of research funding, a constrained institution would not 
increase house staff size regardless of changes in funding, and might reduce house 
staff size only in response to a major change in funding. For example, even if funding 



House staff n 
size 




Maximum sizes of 
house staff at 
each level of 
patient availability 



Availability of 
teaching patients 



Fig. 1— Patient availability constraints on house staff size 



19 



changes caused desired house staff size to shift back and forth between H B and H c , 
the institution at point B in Fig. 1 would not make any change in house staff size; 
the institution would respond only to a change in funding sufficient to make the 
desired house staff size smaller than level H B . 

In principle, patient availability constraints are a short term phenomenon: Over 
time, staff size might be expanded by extending affiliation arrangements. For exam- 
ple, the institution in Fig. 1 might eventually achieve level H c by expanding patient 
availability and moving to point D. In reality, however, patient availability con- 
straints may also prove to be a longer-term phenomenon. Many institutions have 
largely exhausted their locally available affiliation opportunities and many affiliat- 
ed hospitals are experiencing declining patient loads. In some specialties (e.g., OB/ 
GYN) declining patient availability is a trend expected to continue for some time 
into the future. 2 

The present study does not address longer-term responsiveness to research fund- 
ing. To do so, we would require more detailed time-series data on affiliation behavior 
than are available to us. In any case, our primary interest is in whether high 
research funding causes larger house staff size than would otherwise be found, given 
patient availability. The analysis examines research funding effects among cross- 
section samples of institutions, each of which faces a currently fixed availability of 
patients and some of which may face current constraints on house staff size because 
of patient availability. 

In our cross-section data, we observe many different combinations of house staff 
size and patient availability. As shown below, the data consistently reveal a ten- 
dency for house staff size to be larger in institutions with larger patient loads, but 
the data do not provide a means of determining whether any of the institutions in 
our samples have been constrained by patient availability. Therefore, our estimates 
of the effects of research funding may be based on behavior of an unknown mixture 
of constrained and unconstrained institutions. 

A consequence of using a mixed sample is that the results might underestimate 
funding effects in unconstrained departments and overestimate funding effects in 
constrained departments. The results are best interpreted as a general statement 
about "average" behavior for the nationwide set of institutions. However, even this 
interpretation would be relevant only to the particular mix of constrained and 
unconstrained institutions reflected in our data. If patient availability does impose 
constraints on house staff size, the number of institutions facing constraints can be 
expected to increase over the next several years, and the "average" responsiveness 
to such factors as research funding would be expected to decline from levels that 
might be observed in our data. 

Data Sources 

Measures of house staff size and patient care activities by department were 
obtained from the American Medical Association, Directory of Approved Internships 
and Residencies (Green Book). The time required to prepare these data (as described 
below) necessitated a decision to focus on a single clinical department for analysis. 
We have chosen to analyze medicine because of its importance in the clinical train- 

2 It is observed that major teaching hospitals have been increasing their share of the market-wide 
availability of beds. This also is a trend that cannot continue indefinitely. 



20 



ing curriculum and the fairly close correspondence in scope of activities between the 
hospital service (internal medicine) and the medical school department. 

To allow for analysis of several recent years of data, we obtained data from the 
Green Books for academic years 1967-68 through 1969-70, but excluding 1970-71 
because the Green Book was not published that year. The sampling frame for the 
analysis consisted of medical schools in the list of Medical School Affiliations from 
each issue of the Green Book. For each such school, we prepared a list of hospitals 
with major affiliations (as defined by the Green Book) in each year as indicated in 
the Consolidated List of Hospitals. Then we obtained data on the annual admissions, 
average daily census, and annual out-patient visits for internal medicine in each 
affiliated hospital from the List of Approved Residencies. The last also provided data 
on the total number of residency positions available in the hospital, while the List 
of Approved Internships provided the total number of internships available. Finally, 
each of these variables was summed over all the hospitals affiliated with each school. 
This process yielded six yearly samples, each covering roughly 70 to 80 departments 
of medicine. 

Note that the numbers of residencies or internships refer to positions available 
for the year addressed by each Green Book, whereas the patient load data reflect 
conditions in each hospital two years earlier. (For example, admissions data from 
1970-71 would be obtained during 1971-72 for publication of the 1972-73 issue of the 
Green Book.) It is assumed here that the patient load data reported in a particular 
issue are those that were relevant to the choice of the number of house staff positions 
offered in that issue. Further, we assume that each school and hospital in the sample 
expected to fill all available positions. 3 

Data on faculty size in departments of medicine were obtained from the Informa- 
tion Profile System of the Association of American Medical Colleges. These data 
were available by category of faculty (e.g., full-time, associate professors) but could 
be matched with Green Book data only for academic years 1971-72, 1972-73, and 
1973-74. NIH research funding data were obtained, by category of grant, from the 
IMPAC file. 4 If a grant was made to a hospital, we included the funding data for the 
school with which the hospital was currently affiliated. These funding data are 
available for all the years for which house staff size data were collected. 

Of the three data sources, the Green Book seems most likely to contain reporting 
or coding errors; to eliminate the most egregious, we calculated average lengths of 
stay (based on admissions and daily census data) and omitted any observation that 
yielded values greater than 30. Moreover, we examined the time-series of data for 
each school and deleted observations where the value of a variable was not within 
a plausible range of variation. Similar cleaning procedures were applied to the 
remaining data sources, but always with due caution to avoid biasing results in the 
research. 

3 The National Intern and Resident Matching Program reports that approximately 72 percent of all 
available positions in 1975 were filled by the program and has informed us that this has been typical of 
earlier years; hospitals with major affiliations (such as those in our sample) have a better-than-average 
chance of filling positions through the program, and some additional positions are filled by nonpartici- 
pants. 

4 We used the grant description to assign funding to departments by matching each grant title with 
department titles and codes in the Faculty Roster File. The file is maintained by the Association of 
American Medical Colleges. Faculty data are collected by surveying individual institutions, and the 
AAMC assigns department codes to the department titles submitted in the surveys. 



21 



Problems with Proxies 

The data used in the analysis do not necessarily reflect the variables most 
desired for study. In some cases, the analysis was modified to accept the kinds of 
variables on which data were available, while in other cases we assume that the data 
provide approximate measures (proxies) for the variables of interest. For example, 
although we assume that current house staff size reflects expectations about the 
availability of teaching patients (among other factors), our data measure the total 
availability of patients in a previous period. In the analysis we assume that previous 
patient loads are a measure of current expectations and that the availability of 
teaching patients is approximately proportional to the total availability of patients. 

The measures of house staff size pose somewhat more difficult problems. First, 
the designation of internships and residencies may vary among schools and over 
time; in at least some cases, internships and residencies may not differ in any 
substantive way. Therefore, we consider the hypothesis that total house staff size, 
rather than internships and residencies separately, is the variable of interest. 

A second problem is that the data used here measure positions to be filled rather 
than total house staff size though the analysis is concerned with determination of 
the latter. We treat positions offered as a proxy for program size, according to the 
following reasoning: 5 In each year, the number of available positions reflects re- 
placement of losses due to attrition or completion of training, plus a factor (positive 
or negative) reflecting the desired change in overall program size. If the desired 
overall change is zero, then number of positions offered is a proxy for program size. 
If most schools are at their desired program size in a given year, schools with larger 
numbers of offered positions are those with larger program sizes. If most schools 
were, say, expanding by roughly the same proportions, then positions offered would 
continue to reflect cross-sectional differences in program size. However, in compari- 
son with years in which most schools were not expanding, positions offered would 
be larger, across the board. 

Finally, measures of total faculty size may not be appropriate proxies for the 
variables of interest here. We might suppose that house staff size is affected by 
faculty time available for the joint activities, teaching and patient care. Total faculty 
time, as measured by faculty size, includes time devoted to other activities — re- 
search in particular. Thus, faculty size variables may act as proxies for research 
funding as well as for the variable (time for patient care and teaching) we wish to 
measure. The implications of this for the empirical analysis are discussed in more 
detail in the section on empirical methodology, below, and in Appendix A to this 
section. 

Patterns and Trends 

Table 5 presents the means and standard deviations of house staff positions and 
patient load data among medical schools for the six annual samples. Certain pat- 
terns are apparent in the data, though it should be emphasized that none of the 
variable means differ in a statistically significant manner among the six samples. 

5 We have considered alternative hypotheses about the relationship between positions available and 
program size. In particular, we tried alternative specifications reflecting the hypotheses that positions 
available include replacement of losses from the program in the preceding year plus the desired change 
(positive or negative) in current program size. The resulting specifications did not yield new information, 
so in the results presented here, positions available are used simply as direct proxies for house staff size. 



22 



Table 5 

MEANS OF HOUSE STAFF SIZE AND PATIENT LOADS, SIX ANNUAL SAMPLES 3 











YEARS 








COMBINED 


COMBINED 


VARIABLE 




1967-1968 

42.8 


1968-1969 
45.2 


1969-1970 


1971-1972 


1972-1973 


1973-1974 


1967-1970 


197i-1974 


RESIDENCIES 


45.4 


58.2 


55.8 


57.8 


44.5 


57.2 


(RES) 




(27.5) 


(27.8) 


(27.7) 


(39.2) 


(37.5) 


(35.3) 


(27.7) 


(37.4) 


INTERNSHIPS 




38.1 


36.3 


38.1 


42.4 


38.9 


36.7 


37.5 


39.3 


(INT) 




(27.4) 


(24.0) 


(25.4) 


(27.9) 


(22.5) 


(23.1) 


(25.7) 


(24.7) 


ANNUAL ADMISSIONS 11 


8841 


8520 


8458 


10,602 


10,592 


10,877 


8605 


10,687 


< \A) 




(6012) 


(5178) 


(5663) 


(7783) 


(6816) 


(6563) 


(5624) 


(7070) 


AVERAGE DAILY 


CENSUS ' 


379 


383 


352 


388 


403 


372 


• 371 


388 


(ADC) 




(246) 


(243) 


(239) 


(276) 


(269) 


(252) 


(244) 


(266) 


OUTPATIENT VLSI TS b 


52,044 


48,910 


43,503 


57,214 


60,419 


71,041 


48,163 


62,787 


(OPV) 




(40,497) 


(42,919) 


(36,847) 


(60,131) 


(49,028) 


(59,492) 


(40,358) 


(56,546) 


AVKRACI l-EXt.l 


b 
i OF STAY 


15.7 


16.4 


15.2 


13.9 


14.1 


12.6 


15.8 


13.5 


(AES) 




(4.4) 


(4.1) 


(3.8) 


(3.5) 


(4.3) 


(2.9) 


(4.1) 


(3.7) 






.006 


.006 


.006 


.006 


.006 


.006 


.006 


.006 






(.003) 


(.004) 


(.003) 


(.003) 


(.003) 


(.003) 


(.003) 


(.003) 


RES/ADC 




.14 


.15 


.16 


.17 


.17 


.18 


.15 


.17 






(.09) 


(.09) 


(.09) 


(.08) 


(.10) 


(.08) 


(.09) 


(.09) 


INT/AA 




.005 


.005 


.005 


.005 


.004 


.004 


.005 


.004 






(.003) 


(.004) 


(.003) 


(.003) 


(.003) 


(.002) 


(.003) 


(.003) 


INT/ADC 




.13 


. 12 


.13 


.13 


.13 


.11 


.13 


.12 






(.11) 


(.09) 


(.08) 


(.08) 


(.11) 


(.06) 


(.09) 


(.09) 


SAMPLE SIZE 




69 


72 


69 


72 


78 


71 


210 


221 



Standard deviations in parentheses. 

Patient load measures for Internal Medicine Services of hospitals having major affiliations with each school. 

Vhen divided by 365, yields a measure of house staff per patient day. 



The patterns in the data are most apparent when the first three years are 
compared with the last three. Between the two periods, there seem to have been 
increases in numbers of residencies, in annual admissions, in outpatient visits, and 
in residencies per patient day (as indicated by the ratio of residencies to average 
daily census). Average length of stay seems to have declined between the two peri- 
ods; this may reflect a changing patient composition either within hospitals or 
through changes in the nature of affiliation arrangements. 

Aside from the general features of the two periods, there are noticeable peaks 
for 1971-72 in residencies, internships, and annual admissions; and there is a sizable 
step up in annual admissions, average daily census, and out-patient visits between 
1969-70 and 1971-72. Since the data measure available house staff positions rather 
than total house staff levels, the peaks for residencies and internships may indicate 
that schools were moving to a higher level of house staffing in 1971-72. A second 
upward movement in residencies appears to have occurred in 1973-74. 

Since the patient load data are lagged (as noted above), the higher patient loads 
listed for 1971-72 and 1973-74 could have been responsible for an increase in house 
staff positions. Notably, the number of residencies and internships available per 
patient episode (residencies or internships per annual admission) do not vary much 
between 1971-72 and the other years. These observations are consistent with the 
hypothesis that house staff size is importantly affected by the availability of patients. 

Table 6 lists average NIH funding by category of grant for the same annual 
samples described by Table 5. For funding, there are few cases in which the two 
three-year periods exhibit distinctive patterns; instead, there are fluctuations 
throughout the six years. Research, training, fellowships, and career development 
grants seem to have declined substantially in 1972-73, but grants labeled "Other" 
more than compensated, yielding an overall grant level second only to that for 



23 



Table 6 
MEANS OF NIH RESEARCH FUNDING VARIABLES, SIX ANNUAL SAMPLES 3 









YEARS 












GRANT 














COMBINED 


COMBINED 


CATEGORY 


1967-1968 


1968-1969 


1969-1970 


.971-1972 


1972-1973 


1973-1974 


1967-1970 


1971-1974 


RESEARCH 


S636.6 


$575.4 


$545.2 


$635.1 


$577.7 


$746.2 


$583.6 


5650.5 




(645.1) 


(568.7) 


(487.8) 


(551.0) 


(517.5) 


(671.1) 


(302.5) 


(585.6) 


TRAINING 


333.0 


332.6 


311.2 


287.6 


218.8 


311.2 


325.7 


270.9 




(280.3) 


(306.4) 


(318.7) 


(238.2) 


(237.0) 


(353.8) 


(302.5) 


(283.0) 


FELLOWSHIPS 


29.4 


26.4 


17.3 


15.9 


9.1 


29.2 


24.4 


17.8 




(56.9) 


(94.8) 


(41.5) 


(28.2) 


(17.7) 


(47.2) 


(91.7) 


(34.0) 


CAREER DEVELOPMENT 


69.4 


71.8 


72.7 


62.0 


57.3 


58.5 


71.3 


59.2 




(81.7) 


(94.8) 


(97.5) 


(63.5) 


(65.3) 


(63.7) 


(61.3) 


(64.2) 


I'ROCRAM PROJECT 


330.0 


316. 1 


302.6 


389.2 


492.0 


636.9 


316.2 


505.1 




(628.7) 


(611.4) 


(315.2) 


(661.7) 


(871.3) 


(969.6 


(321.3) 


(850.1) 


CL1XTCAI. RESEARCH CENTER 


171.2 


153.9 


148.3 


137.6 


140.1 


158.5 


157.7 


145.2 




(333.7) 


■ (314.4) 


(315.2) 


(273.0) 


(287.2) 


(293.7) 


(321.3) 


(284.9) 


OTHER 


• 


— 


— 


— 


223.5 
(371.0) 


— 


— 


78.9 
(244.9) 


TOTAL 


1.569.5 


1,476.3 


1,397.3 


1,527.4 


1,718.5 


1,940.6 


1,481.0 


1,727.6 




(1,739.2) 


(1,706.0) 


(1.620.4) 


(1,286.6) 


(1,803.5) 


(2,007.4) 


(1,691.0) 


(1,734.8) 


SAMPLE SIZE: 


69 






72 


78 


71 




221 



Standard Deviations in parentheses. 



1973-74. Average NIH funding over all categories reached its lowest point in 1969-70 
for departments in our samples. 

The initial analysis of house staff size, reported below, uses the data to which 
Tables 5 and 6 refer. To include faculty data, the sample was reduced to years 
1971-72, 1972-73, and 1973-74. Table 7 presents descriptive data concerning the 
resulting sample. 



Table 7 
MEANS OF HOUSE STAFF AND FACULTY SIZE: THREE ANNUAL SAMPLES''' 











COMBINED 


VARIABLE 


1971-1972 


1972-1973 


1973-1974 


YEARS 


HOUSE STAFF: 










RESIDENTS 


58.2 


55.8 


57. S 


57.2 




(39.2) 


(37.5) 


(35.3) 


(37.4) 


INTERNS 


42.4 


38.9 


36.7 


39.3 




(27.9) 


(22.5) 


(23.1) 


(24.7) 


b 
FACULTY: FULL-TIME 


54.4 


55.9 


60.1 


56.7 


(FACET) 


(35.0 


(39.6) 


(40.1) 


(38.4) 


PART-TIME (FACPT) 


12.2 


9.2 


8.6 


10.0 




(18.1) 


(9.3) 


(9.3) 


(13.0) 


VOLUNTEER (EACVOL) 


139.2 


122.6 


111.9 


124.6 




(120.0) 


(85.8) 


(78.9) 


(97.0) 


SAMPLE SIZE: 


72 


78 


71 


221 



Standard deviations in parentheses. 
Faculties of departments of Medicine. 



24 



Since the same annual samples are used in constructing all the tables, the means 
and standard deviations of the house staff variables in Table 7 do not differ from the 
corresponding values in Table 5. The faculty size variables suggest that numbers of 
part-time and volunteer faculty may have declined slightly between 1971-72 and 
1973-74; average full-time faculty size appears to have increased somewhat. These 
results, together with those in the earlier tables, suggest that there was neither 
substantial growth nor substantial decline in teaching, patient care, or research 
activities associated with major teaching hospitals over the period 1971-1974. 

Further information about the data is yielded by examining plots of house staff 
size against measures of patient availability and against NIH research funding. 
Figure 2 plots the sum of residencies and internships available in each institution 
against a measure of patient availability that reflects total inpatient days but 
weights days near the. end of an average episode less heavily than days near the 
beginning. 6 Figure 3 plots the same measure of house staff size against total NIH 
research funding for each institution. Both figures use data for the academic year 
1971-72 (patient data lagged as noted above), but the plots are representative of those 
for other years. 



House 290 
staff 



3,500 •■- 
24 



91,000 



Patient 
availability 



Fig. 2— Plot of data for academic year 1971-72: House staff 
size and patient availability 



In Fig. 2, the observations occupy a fairly narrow band, suggesting that house 
staff size tends to be larger in institutions with larger patient availability. There also 
appears to be substantial variation in house staff size among institutions with very 
similar patient availability. At face value, the figure seems to indicate that there is 
considerable discretion in setting house staff size and that few — if any — institutions 
are operating on the kind of constraint line postulated in Fig. 1, above. However, 



" The measure is actually defined as the product of annual admissions and the natural logarithm of 
average length of stay. The conceptual basis for the measure is described below. 



25 



House staff 
size 



290 



522,000 



24 



$5,000,000 

NIH research funding 



Fig. 3 — Plot of data for academic year 1971-72: 
House staff size and NIH research funding 



our measures of house staff size and patient availability are sufficiently crude that 
we cannot be sure that the "scattering" of points in the figure does not merely 
represent error. For the same reason, we do not use the data to speculate about 
which institutions are actually constrained. 

In contrast, Fig. 3 shows no obvious relationship between house staff size and the 
more accurate measure of NIH funding. Although institutions do cluster in the 
lower left-hand quadrant of the figure, every quadrant contains some observations. 

Neither Fig. 2 nor Fig. 3 addresses the possibility that research funding affects 
house staff size once patient availability is taken into account. To achieve this 
objective, we turn to multivariate statistical techniques that permit us to relate 
house staff size simultaneously to both patient loads and research funding. 



Empirical Technique 

If we used regression analysis to relate house staff size to patient availability, 
the regression would fit a line through the points illustrated in Fig. 2. Multiple 
regression allows us to include research funding as an additional explanatory varia- 
ble. Heuristically, the estimated coefficients of research funding describe whether 
the vertical distances between observations and the line are systematically related 
to levels of research funding. For example, a positive coefficient on research funding 
would mean that an institution with less than average funding would tend to lie 
below the line, and an institution with higher than average funding would tend to 
lie above the line. 

The results would not tell us, however, whether research funding causes an 
institution to lie on or off the line. An institution may lie on or off the line for a 
variety of reasons, and it may be true that some other factor that is simply correlated 
with research funding is actually responsible for variations in house staff size. 

Faculty size might be one such factor. We know that faculty tends to be larger 



26 



in institutions with higher research funding. 7 It is plausible also to suppose that 
house staff size might tend to be higher (even relative to patient loads) in institutions 
that have more faculty. If so, faculty size should be included in the equations as a 
control variable so that the funding coefficient would measure only the effect of 
funding net of faculty size effects. However, there are models of house staff determi- 
nation that suggest that faculty size should be omitted from the analysis. In particu- 
lar, if: (1) only the portion of faculty time devoted to teaching and patient care is 
relevant to house staff size, and (2) faculty time for teaching and patient care is 
independent of research funding, then including total faculty size can yield biased 
estimates of funding effects. 

Appendix A to this section describes several alternative models of house staff 
determination and lists the empirical implications of including or excluding faculty 
size variables under each model. This analysis suggests that each equation should 
be run both with faculty variables and without them. The differences in results 
between the two specifications can be used to generate additional tentative infer- 
ences about house staff determination and the role of research funding. 

Analysis and Results 

Setting aside, for the moment, the effects of faculty size and research funding, 
consider the specification of a relationship between patient load and house staff size. 
Our data provide a crude means of describing both the duration of patient care 
(average length of stay) and the number of patients treated. Although there may be 
differences in diagnostic case mix or complexity among hospitals, we have no means 
of measuring such differences except through their influence on lengths of stay. 

To specify the relationship, we initially considered two hypotheses: First, we 
postulated that each resident or intern spends a certain amount of time, on average, 
with each patient for each postadmission day of stay; according to this hypothesis, 
house staff input is a simple function of total inpatient days. Second, we postulated 
that the amount of house staff involvement with a patient declines continuously 
(though perhaps slowly) from the first day of stay to the last. A simple means of 
specifying the second hypothesis is to form a measure of patient days that weights 
days near the end of an episode less heavily than days near the beginning. One such 
measure is given by the product of annual admissions and the natural logarithm of 
average length of stay; this measure was used in Fig. 2, above. A disadvantage of 
the measure is that it does not offer a very satisfactory means of testing the hypoth- 
esis on which it is based. 

A second specification is derived as follows: The amount of house staff time 
devoted to each patient is the sum of the amounts of involvement on each day of an 
inpatient episode. Using a straight line to describe how daily involvement declines 
over the length of stay, the basic equation is: 

HOUSE STAFF INPUT ON iTH DAY = a + y3 (iTH DAY OF STAY) (1) 

where P is expected to be negative to reflect declining daily inputs. Therefore, the 

7 Evidence on this point is shown in another study in this report. A further variable of potential 
importance is the number of graduate fellows in an institution. Data on this group are unavailable. 



27 



total input for an episode with a given length of stay (LOS) is the integral of Eq. (1) 
from zero to the value of LOS: 

HOUSE STAFF INPUT PER EPISODE = a LOS + £ (LOS 2 /2) (2) 

where /3 is still expected to be negative. To get the total house staff input over all 
patients, Eq. (2) should be summed over all patients. Since we have data only on 
average lengths of stay and annual admissions, we must approximate the desired 
sum by multiplying the right-hand side of Eq. (2) by the total number of admissions 
per year (AA). 

The final specification of inpatient load includes two variables, the first of which 
is annual admissions multiplied by average length of stay (AA x LOS); this variable 
is equivalent to total annual patient days. If house staff daily input is constant over 
the length of stay, then only the first variable would have a nonzero coefficient. A 
test of the hypothesis that staffinputs decline during an inpatient episode is whether 
the estimated coefficient of the second variable (LOS 2 /2 x AA) is negative; the 
coefficient of the second variable is equivalent (by derivation) to the (i in Eq. (1). 

To complete the specification of patient loads, we include outpatient visits (OPV) 
as an additional explanatory variable. Then we add funding variables to the equa- 
tions. There were too few observations in each sample to include separate variables 
for all the categories on NIH funding. Therefore, in preliminary analysis we used 
subsets of the categories in alternative specifications, hoping to determine whether 
any particular categories of funding are especially relevant to determining house 
staff size. Since there was no clear evidence that certain kinds of funding are particu- 
larly relevant, the results shown here use only total NIH funding, FUND. 

The equations to be estimated describe residencies (RES), internships (INT), and 
total house staff positions (HSTAFF = RES + INT) as follows: 



ie start positions (HSTAFF = KUS + INT) as follows: 

RES = INTERCEPTr + a R (LOS x AA) + £ R (LOS 2 /2 X AA) + 

y R OPV + 8rFUND + £ R , 
INT = INTERCEPT, + a,(LOS x AA) + £,(LOS 2 /2 x AA) + 

7[OPV + S,FUND + e If 
HSTAFF = INTERCEPTh + a H (LOS X AA) + H (LOS 2 /2 X AA) + 

y H OPV + ShFUND + e H . 



(3) 
(4) 
(5) 



We expect the estimates of a and y to be positive and the estimates of P to be 
negative or zero. The error terms (e R , e h e H ) reflect random factors affecting house 
staff size, possibly including faculty size. The INTERCEPT terms would measure the 
average effects of any omitted variables. 

Results from this analysis are presented in Appendix B, Table B.l; only coeffi- 
cients that are statistically different from zero with at least 95 percent confidence 
are reported. The coefficients in the HSTAFF equations are equal (with some round- 
ing error) to the sum of the corresponding coefficients in the RES and INT equations. 
In general, the RES and INT equations perform quite well individually and differ 
in their coefficient estimates. For these reasons, we rely primarily on the separate 
RES and INT equations in the remainder of this discussion. 

Coefficients of determination (R 2 , corrected for degrees of freedom) measure the 
proportion of variance (i.e., dispersion) in the dependent variable that is "explained" 
by the variables in the equation. For RES, patient load and funding variables to- 
gether explain 61 to 72 percent of the variation in house staff size among institu- 



28 



tions; for INT, the equations explain 42 to 68 percent. These figures are unusually 
high for any cross-section analysis, and are surprisingly high given the crudeness 
of the data and the simplicity of the models used here. 

In all cases for which the patient load variables yield statistically significant 
coefficients, the signs are those we expected. Interestingly, outpatient visits rarely 
yield significant coefficients in the RES equations; this may reflect a larger role of 
residents in outpatient care. 

As noted above, a significant negative estimate for P may be interpreted as 
evidence that house staff input declines during the length of stay. Obtaining a 
significant coefficient is difficult because we had to use average lengths of stay for 
each institution, because there is relatively little variation in length of stay among 
institutions in the sample (particularly in 1973-74, as shown in Table 5), and because 
the variable used to estimate (i is highly correlated with the variable used to esti- 
mate a. 8 Nevertheless, the estimate of/? manages to be significant in four of the 
INT equations and two of the RES equations. The results seem to support the 
hypothesis that house staff input declines during an inpatient episode. 

The coefficients of the patient load variables appear to differ substantially from 
year to year. We do not offer any particular interpretation of this fact. The measures 
of house staff used here are not necessarily comparable from year to year (as noted 
above) and none of the coefficient estimates differ from year to year in a statistically 
significant manner. 

The funding coefficients are almost never statistically significant. Even when 
they are significant (for RES in 1968-69 and 1969-70), the estimated values are very 
small. For example, for 1969-70 the results show that funding would have had to 
differ by over $300,000 to generate a difference of one in the number of residents. 
At the sample means for that year, the results imply that a difference of 24 percent 
in NIH funding might have generated only little more than a 2 percent difference 
in residencies. 

Now let us consider how the funding effect estimates are affected by including 
faculty size variables in the analysis. This can be done only for the last three annual 
samples. To simplify the analysis, we obtained only one coefficient for inpatient 
loads by using the logarithmic patient load variable described above; preliminary 
analyses showed that similar results are obtained using the pair of inpatient vari- 
ables. Although preliminary analyses used several different measures of faculty size, 
the measure used here performs as well as any of the combinations of separate 
variables we tried. The measure is FAC = full-time faculty plus part-time faculty 
plus 1/10 of volunteers. 

For purposes of comparison, Table B.2 in Appendix B presents pairs of otherwise 
identical equations for each of the three annual samples for which faculty data are 
available; the only d'fference within a pair of equations is that FAC is first excluded 
and then included. The equations for interns differ from those for residents in that 
OPV is omitted; similar results are obtained when OPV is included, and the coeffi- 
cients of OPV are almost never statistically significant. 

Except for the intercept terms, the coefficients in residents equations are strik- 
ingly similar among the three years; considerable similarity is also observed for the 
interns equations. Therefore, it seems reasonable to combine the three years of data 

8 A regression cannot obtain a precise estimate of the coefficient of a variable unless the variable varies 
in the sample and varies independently of other variables in the equation. 



29 



in equations that include year dummy variables (YR73 and YR74) to allow the 
intercepts to vary by year. Results for the combined years are: 

RES = 10.29 — 3.72 YR73 — 2.93 YR74 
(3.53)** (3.58) (3.70) 



+ .0014 AA • LOG(ALS) + .00012 OPV 
(.0001)** (.00003)** 



+ .0013 FUND R* = .66 

(. 0006) 



(6) 



(7) 



(8) 



RES = 3.44 — 2.55 YR73 — 1.63 YR74 
(3.65) (3.42) (3.54) 



+ .0013 AA ■ LOG(ALS) + .00012 OPV 
(.0001)** (.00003)** 



- .0013 FUND + .186 FAC R z = .69 

(.0010) (.039)** 



INT = 11.81 - 4.20 YR73 - 6.24 YR74 
(2.45)** (2.49)* (2.56)** 

+ .0028 AA • LOG(ALS) + .0024 FUND R 2 = .62 
(.0001)** (.0006) 



INT = 7.91 - 3.54 YR73 - 5.55 YR74 
(2.58)** (2.42) (2.48)** 



+ .0010 AA • LOG(ALS) - .0005 FUND 
(.0001)** (.0007) 

+ .105 FAC . R 2 = .64 

(.028)* 

In these equations, the figures in parentheses are standard errors, one asterisk 
denotes a coefficient that is statistically nonzero with at least 90 percent confidence; 
two asterisks denote 95 percent confidence. Funding is measured in thousands of 
dollars. 

The consistency of the negative signs on YR73 and YR74 suggusts that residen- 
cies and internships were fewer after 1972, net of effects of changing patient loads, 
faculty size, and funding. If so, 1972 may have been a year of expansion in house 
staff size. 

The patient load variables always yield significant coefficients with the expected 
signs. These variables alone explain over 50 percent of the variance in house staff 
size. 



(9) 



30 



When included, the faculty size coefficients are statistically significant and posi- 
tive. At the means of the variables, the coefficients imply that a 10 percent difference 
in faculty size implies about a 3 percent difference in residencies and about a 2 
percent difference in internships among institutions. 

The coefficients of funding are statistically nonzero with better than 80 percent 
confidence in both of the residency equations — but change sign from positive to 
negative when faculty size is included. For interns, the funding coefficient is signifi- 
cantly nonzero at 80 percent confidence only when faculty size is excluded. 

Referring to the chart in Appendix A, the results are clearly inconsistent with 
models II, IV, V, and VIII. The results for residencies also seem to rule out model 
I. The three remaining models (III, VI, and VII) all suggest that house staff size is 
based not on total faculty size but on faculty time available for teaching and patient 
care. Moreover, the plausible models do not suggest that faculty time available for 
patient care and teaching declines with research funding; two of the models assume 
that available time for teaching is higher when research funding is high (perhaps 
suggesting that faculty expands by more than the amount directly involved in 
research). 

One of the three plausible models suggests that despite apparently nonzero 
coefficients, funding has no effect on house staff size. The other two models suggest 
that the estimates are an upper bound on the true effect. Assuming that our results 
are an upper bound, we find that any direct effects of funding are at most very small: 
a difference of one million dollars would imply at most a difference of about 1.2 in 
residencies and a little less than one in internships. 

Summary of Results 

The analysis presented here possesses several shortcomings, most due to lack of 
detailed, accurate data on the variables of interest. Nevertheless, the data yield 
plausible and statistically meaningful results when applied to the models used in 
this study. We find that patient loads explain much of the variation among institu- 
tions in house staff size and that the data reveal a plausible decline in daily house 
staff input during an episode of inpatient care. 

In general, the results do not show strong statistical significance for coefficients 
of research funding and the coefficients are very small. The results simply yield no 
evidence of a strong or consistent direct effect of research funding on house staffsize. 
The results also fail to show that research competes with teaching for faculty time. 

As we argued above, the reason that research funding effects appear to be so 
weak may be that patient-availability constraints overwhelm research funding in 
affecting decisions with regard to house staff size. If so, and if the constraints con- 
tinue in the future, prospective research funding policies may be expected to have 
little effect on house staff size. 



APPENDIX A 

The appropriateness of including or excluding faculty size measures as control 
variables in house staff regressions depends greatly on our model of house staff 
determination. Specifically, depending on the model, the coefficient of research fund- 
ing might be biased either by including or by excluding faculty. Here we show that 



31 



pairs of results with faculty included and excluded can help distinguish among 
alternative models. This is shown by working through some simple deterministic 
models and then by summarizing some additional models by means of a simple 
chart. 

Consider a simple deterministic model: 

H = a,P + b,F, (i) 

where H = house staff size, P = patient load, F = faculty size, and both a x and bi 
are positive parameters. The model says that research funding does not affect house 
staff size, but faculty size does. However, suppose that there is a relationship be- 
tween faculty size and research funding, R: 

F = c, + (UR, (ii) 

where Ci and d! are positive parameters. 

If we related house staff to patient loads and research funding, we would observe 
the relationship: 

H = aP + b"R + c" (iii) 

where a, b, and c are estimated parameters. By substituting Eq. (ii) in Eq. (i) we find 
that: 

H = a L P + b, F 

= ajP + b t (Ci + d x R) 

= ajP + b!d + bidjR. (iv) 

Therefore, in our estimates of Eq. (iii), we would find that: 

a = a u b = b x di and c = bjCj. (v) 

The results would yield a positive coefficient (b) for research funding even if research 
funding has no effect on house staff size. Complex, stochastic models based on the 
same relationships as the foregoing equations yield the same kind of result: Omit- 
ting faculty size might generate biased estimates of the effects of research funding. 
If the preceding model were essentially correct, the problem could be corrected 
by including faculty size as an additional variable. If we estimated: 

/\ s\ x\ /\ 

H = aP + bF + fR + c, (vi) _ 

we would correctly obtain a zero value for f, where a, b, f, and c are estimated 
coefficients. 

However, if the foregoing model is not correct, including faculty size variables 
may not generate appropriate results. For example, suppose that house staff size is 
affected by faculty time available for teaching and patient care activities, T, rather 
than by total faculty time (as measured by F); that is, suppose that: 

H = a 2 P + b 2 T, (vii) 

and 

F = T + S, (viii) 



where S = time spent in research. Suppose further that T and S are determined 
independently (i.e., T and S are determined and these in turn determine total faculty 
size) and only S is affected by research funding: 

S = d 2 R. (ix) 

Substituting (ix) and (viii) into (vii), we have: 



(x) 



H = a 2 P + b 2 T 




= a 2 P + b 2 (F - 


- S) 


= a 2 P + b 2 F - 


- b 2 d 2 R 



Therefore, if we estimated Eq. (vi), we would find that: 

a = a 2 , b = b 2 , and f = — b 2 d 2 . (xi) 

Including faculty size variables would cause the estimated equation to generate a 
negative coefficient for a research funding variable even though the model (Eq. vii) 
says that funding doesn't affect house staff size. In this case, it would be more 
appropriate to omit faculty size from the empirical analysis. 

The characteristic of the second model that is essential to obtain this result is 
the assumption that research funding does not affect the component of faculty time 
that influences house staff size. Alternatively, if T were a function of R (if research 
funding affected the allocation of faculty time to teaching and patient care), then 
we would get a biased estimate of the funding coefficient regardless of whether 
faculty size is included in the analysis. Specifically, suppose: 

T = g 3 R. (xii) 

Then, from the model in Eq. (vii), we have: 



(xiii) 



(xiv) 

If g 3 is positive (T increases with R), then Eq. (xiii) shows that we would get a positive 
coefficient for R when faculty size is omitted and Eq. (xiv) shows that we would get 
a negative coefficient when faculty size is included. Opposite results would be ob- 
tained if g 3 were negative. 

Figure A-l summarizes the implications of various models when faculty size is 
included or excluded from analysis. The chart includes not only the results from the 
foregoing models that assume research has no direct effect on house staff size, but 
also includes results for corresponding models that assume research funding has an 
effect on house staff size. 

Note that the chart describes results for deterministic models. If the models 
were stochastic, empirical results would not be as clear-cut as the chart implies. For 
example, instead of obtaining a zero coefficient in case 2, we might obtain a positive 
coefficient that is not statistically different from zero. 

Of the pairs of results shown in the chart, only the pair for Model number II is 
distinctive; at least two different models could produce any of the other pairs of 



H = 


a 2 P + b 2 T 




= 


a 2 P + b 2 g 3 R 




H - 


a 2 P + b 2 (g 3 (F 


- S)) 


= 


a 2 P + b 2 g 3 F - 


b 2 g 3 d 2 R. 



33 





Modi' 1 


lilttil LV 

Form.it ion 


I k 


Mod..-] 
Number 


Case 

Number 


Intimating Equatio 


ResuJ ling 
Coefficient 
l of Funding 


S talus 


II 


■ V + V 


<n 


(1) 
(2) 


Exclude Faculty 


Posit ive 


Should lu .'.,-<-■■ 




Inc lude Facul ty 


Zero 


Correct result 




■ .1,1' + b,l 


Y = T + S 
S = d,R 

T indi-peni 

of K 


en I 


(IT) 


(J) 


Exclude Faculty 


Zero 


Correct result 




(4) 


Im ludo Fm-ul ty 


Negative 


Should be zero 


II 


F - T + S 

S = d R 

! ■ R,R 




(HI) 


(5) 


Exclude F.u ulLy 


Posit ive 


Should be zero 




(6) 


Enc lude Faculty 


Neg.it i vc 


Should be zero 




F - T + S 
S = d,R 

T = g 4 R 

g 4 < o 




(IV) 


(7) 


F.xclude FacuJ ty 


Negat ive 


Should be zero 




(S) 


Include Faculty 


Positive 


Should be zero 


H 


- V + V 

+ f-R 
y 

h >0 


F - Cl + d 


1 R 


(V) 


(9) 


F.xclude Faculty 


Positive 
(Larger than 

Case (10)) 


Biased up 




(10) 


Include Facul ty 


Posi t ive 


Correct resul t 




= a P + bjT 
+ 11 
f 5 >0 


F = T + S 

S = d,R 

T independent 
of R 


(VI) 


(ID 


Exclude Faculty 


Positive 
(Larger than 
Case (12)) 


Correct result 




(12) 


Include Faculty 


Positive or 
Negative 
(Possibly zero; 


Biased down 


H 


F = T + S 
•S = d,R 

T = gjR 

g 3 > 




(VII) 


(13) 


Exclude Facul ty 


Positive 
(Larger than 
Case (14)) 


Biased up 




(U) 


Include Faculty 


Positive or 
Negative 
(Possibly zero) 


Biased down 




F = T + S 
S = d 2 R 

T - gjR 

gj < 




(VIII) 


(15) 


Exclude Faculty 


Positive or 
Negative 
(Possibly zero) 


Biased down 




(16) 


Include Faculty 


Positive 


Biased up 



Fig. A-l— Chart of theoretical outcomes 



outcomes. Thus, the chart illustrates the difficulty of using the empirical results of 
this study to generate precise conclusions about how house staff size is determined. 
At the same time, the chart shows that when data permit, models should be esti- 
mated with faculty variables both included and excluded; the pair of results so 
obtained can at least focus attention on a small subset of the models in the chart. 



Career Outcomes of Graduates 

In recent years, a broad national concern over a general shortage of physicians 



34 



has given way to more specific concern over shortages of particular types of physi- 
cians, particularly those in the primary care specialties. Since primary care involves 
responsibility for a patient's overall health care — treatment of the "whole patient" 
— it is not surprising that some have inferred that a center's biomedical research 
activities, which are usually highly focused on details of a particular health problem 
or body function, are at odds with a center's primary care training functions. 

The supposed mechanism of a center's influencing students to choose nonpri- 
mary care specialties is the role model of the biomedical research "superstar" on the 
faculty. Because these individuals enjoy prestige in academic medicine, it is pre- 
sumed that medical students seek to model their careers after these specialists. An 
extension of this logic is that the more research intensive an institution or a depart- 
ment, the less likely its graduates are to enter primary care specialties. 

The questions we address in this section are, what influence does a medical 
school's federal research funding have on the professional career paths of its M.D. 
graduates, in terms of both the specialty chosen and the type of practice? For the 
purposes of this section, we define four specialty groups: internal medicine, all other 
primary care specialties (general practice, family practice, pediatrics, obstetrics and 
gynecology), surgical specialties, and other nonprimary care specialties. 9 By type of 
practice, we are referring to whether the physician chooses patient care, administra- 
tion, teaching, or research. 

To investigate these questions, we model these two aspects of career choice as 
a function of federal research funding and individual characteristics of physicians. 

The Data. We have three data files relevant to the questions being considered. 
One is the AMA Master File of Physicians, which contains extensive information 
on postgraduate medical education and practice characteristics of all M.D.s. The 
second file was compiled by Rand in connection with its three-year study of federal 
program efforts. This file contains background data on individual students from ten 
academic medical centers, including age, sex, medical school grade-point average, 
selectivity index for undergraduate (premedical school) college, standardized test 
scores, and other variables. 10 The third data file is the IMPAC file created and 
maintained by the Division of Research Grants of the NIH. This file includes infor- 
mation on the amounts of federal funding to medical schools by department over the 
period 1967 through 1975. 

Our data on practice characteristics of physicians from the AMA master file are 
as of December 1974. Our data on student characteristics in ten medical schools are 
for the classes of 1955, 1960, 1965, 1969, and 1972. Since the funding data cover 1967 
to 1975, we eliminate the 1955 and 1960 classes from our analysis. We thus analyze 
data for the 1965, 1969, and 1972 classes from ten medical schools. 

The Analysis. Our objective is to discover what factors influence the types of 
medical careers that physicians chose. For purposes of this analysis we consider two 
aspects of these careers: (1) the medical specialty (e.g., surgery, pediatrics) in which 
the physician is trained, and (2) the primary activity in which the physician is 

9 Both the specialty of practice and the type of practice are defined by the physician who fills out 
the AMA survey. Hence "internal medicine," for example, does not signify board certification or even 
level of training but rather what the physician has designated'as his type of practice. 

10 A selectivity index of undergraduate schools was developed on the basis of material in the "College 
Admissions Selector." The scale runs from 9 points for colleges with the most competitive admissions 
policies to 1 point for a special category of colleges for which admissions is not based primarily on 
academic criteria. See Barron's Profiles of American Colleges, Barron's Educational Series, Woodbury, 
N.Y., 1972, pp. xxii-xxix. 



35 



engaged (patient care, administration, teaching, or research). We have data on these 
two aspects of the careers of the graduates of sample classes from ten medical 
schools, and we are seeking to determine, after the fact, what factors appear to 
explain the career choices of these graduates. 

Discriminant analysis is an appropriate statistical technique to use in such a 
retrospective analysis. It uses a set of independent or explanatory variables to 
statistically distinguish categories on some dependent variables. The model is some- 
what similar to a regression model except that the dependent variable is categorical. 
Linear combinations of the independent variables, called discriminant functions, 
are formed. The coefficients in these discriminant functions are chosen to maximize 
the separation of the groups: to maximize the differences in the discriminant scores 
(the value of a discriminant function for a particular individual and the sample) of 
the groups. These discriminant functions can be analyzed directly and can also be 
used to estimate the relative probabilities of membership, called the classification 
functions, in each of the groups. (The difference between the discriminant scores of 
two groups equals the natural logarithm of the odds that the individual is from one 
group rather than the other.) 

We use four specialty groups: internal medicine, other primary care specialties, 
surgical specialties, and other nonprimary care specialties. Among the 1965, 1969, 
and 1972 graduates from these ten schools, 20.0 percent are in internal medicine 
(IM), 20.3 percent are in other primary care (OPC) specialties, 25.5 percent in surgi- 
cal specialties (SS), and 34.2 percent in other nonprimary care (ONPC) specialties. 
Among these same graduates, 87.8 percent are in patient care, 2.5 percent in ad- 
ministration, and 9.7 percent are in teaching or research. 

The independent variables we use in the discriminant analysis are shown in 
Table 8. 

Findings and Conclusions. We consider the discriminant analysis results for 
specialty choice, and whether the physician is in patient care or academic medicine. 
For both cases, we distinguish the effects of research funding by department to the 
medical schools where the physicians received their medical degrees, from the 
effects of research funding by department to medical schools and their affiliated 
hospitals where the physicians did their first residencies. 

For all four discriminant analyses, the background characteristics of physicians, 
especially rank in medical school class, are the principal determinants of career 
choice. The research intensity of institutions is of less importance and adds very 
little to our ability to predict career choices. 

In the case of specialty choice, sex and rank are the first two variables to enter 
the discriminant analysis, with proportionately more males choosing internal medi- 
cine and the surgical specialties and with physicians who ranked higher in their 
medical school clasees more frequently choosing internal medicine and to a lesser 
degree the surgical specialties. These results hold both for research funding associat- 
ed with the institution where the physician's undergraduate medical education took 
place and for research funding associated with the institution where the first resi- 
dency took place. See Tables 9 and 10. Funding for the primary care specialties other 
than internal medicine during a physician's undergraduate medical education is 
significantly associated with the eventual choice of one of those specialties, and total 
research funding for all clinical science departments is significantly associated with 
a choice of internal medicine and surgical specialties. Research funding for internal 



36 



Table 8 
THE VARIABLES USED IN THE DISCRIMINANT ANALYSIS 



Variables 



Definitions 



Mean Values 



5. 


40 


552 


.70 


565 


.92 


565 


.53 


555 


,46 


3. 


10 


3, 


,10 



SELINDX Selectivity index of the undergraduate 
college (a 1 to 9 scale) 

MCATV MCAT verbal score 

MCATQ MCAT quantitative score 

MCATS MCAT science score 

MCATG MCAT general score 

GPATOTAL Total grade-point average in medical 
school 

GPASCI Grade-point average for science 

RANK Rank in medical school class, transformed 

to approximate mean and standard deviation 
of 500 and 100, and with higher values 
representing higher rank in class 

Sex Dummy variable for sex (male - 0, female = 1) 

FDIM Federal research funding for internal 

medicine departments 

FDOPC Federal research funding for other primary 
care departments 

FDSS Federal research funding for surgical 

specialties departments 

FDONPC Federal research funding for other 
nonprimarv care departments 

FDCST Total federal funding for clinical science 
departments 

FDBST Total federal funding for basic science 
departments 

The measure of selectivity of undergraduate schools is based on 
the evaluation of Barron's Profiles of American Colleges. We assigned 
a value of 9 to Barron's "Most Competitive" category, 7 to "Highly 
Competitive," 5 to "Very Competitive," and so on. Most schools in the 
country rank 3 or lower on this scale. The mean for the applicant 
groups at the sample medical school ranges between 4.5 and 5.3. See 
Barron's Profiles of American Colleges, Barron's Educational Series, 
Inc., Woodbury, N.Y. , 1972, pp. xxii-xxix. 

Medical College Aptitude Test. 

The funding variables are all average annual funding figures in 
millions of dollars for the period 1967 to 1975. The funding variables 
by departments include research grants, teaching grants, research 
support grants, fellowship grants, career development awards, program 
project grants and clinical research center grants. 



10 


.84 




.07 


3 


.68 




.48 




. 55 




.84 


5 


.78 


2, 


,75 



37 



medicine during a physician's residency is significantly associated with the eventual 
choice of internal medicine as a specialty. 

All of the funding variables together add very little to our ability to predict 
specialty choice in the discriminant analysis. In the case of research funding to the 
medical school (and its affiliated hospitals) where the physician received his M.D., 
the discriminant analysis correctly classified 35.9 percent of the physicians into the 
four specialty groups, and was able to correctly classify 35.6 percent using only the 
background variables and no funding variables. Similarly, for funding to the medi- 
cal school and its affiliated hospitals where the first residency was taken, the dis- 
criminant analysis correctly classified 36.3 percent (and the same 35.6 percent with- 
out the funding variables). We would expect to classify 25 percent correctly with 
random assignment. 

For the choice between academic medicine and patient care, the first two vari- 
ables to enter the discriminant analysis for both cases (i.e., funding to undergradu- 
ate medical institution and funding to institution associated with the physician's 
first residency) are RANK and MCATQ. Higher rank in class and higher MCAT 
quantitative scores are both associated with a choice of medical education or re- 
search careers. These two variables are the only variables in the analysis significant- 
ly related to the choice between academic medicine and patient care when we 
consider federal research funding to the physician's undergraduate medical institu- 
tion. One other variable, research funding for the surgical specialties, is significantly 
related to this choice when we consider funding to the institution associated with 



Table 9 

DISCRIMINANT ANALYSIS RESULTS FOR EFFECT ON TYPE OF PRACTICE 3 OF FEDERAL 
RESEARCH FUNDING TO THE INSTITUTION WHERE THE M.D. WAS TAKEN 





S 


ignif icance 


Classi 


ficatlon Functior 


i Coefficients 




Patient 




Medical Teaching 


Variables 




Level d 


Care 


Administration 


or Research 


RANK 




<.001 


.044 


.045 


.050 


MCATQ 




.019 


.067 


.063 


.072 


Sex 




.109 


-1.781 


-2.152 


-.547 


MCATV 




.172 


.064 


.070 


.066 


Constant 






-46.508 


-48.407 


-53.549 


(Other varia 


bl 


es 








not entered) 


e 











The F-ratio for the difference between groups as distinguished 
by the discriminant analysis is significant at the .001 level for the 
difference between patient care and teaching plus research. The 
F-ratios for the differences between administration and either of the 
other two practice types are not significant at the .05 level. 

Differences between the coefficients across groups for a given 
variable represent the estimated effect of that variable on the 
relative probabilities of membership in those groups. 

The variables are listed in the order in which they entered 
the discriminant analysis. 

The significance level for each variable is the statistical 
significance of that variable's contribution to separating the groups. 

No funding variables enter this discriminant analysis. 



38 



Table 10 

DISCRIMINANT ANALYSIS RESULTS FOR EFFECT ON TYPE OF PRACTICE 3 OF FEDERAL 
RESEARCH FUNDING TO THE INSTITUTION WHERE THE FIRST RESIDENCY WAS TAKEN 





Significance 
level d 




Classification Func 


tion 


Coefficients 


c 
Variables 


Patient 
Care 


Administration 


Med 


ical Teaching 
or Research 


RANK 


<.001 


.044 


.044 




.048 


MCATQ 


.019 


.070 


.066 




.076 


FDSS 


.035 


1.264 


1.351 




2.010 


SEX 


.062 


-1.632 


-1.827 




-.096 


FDCST 


.202 


-.140 


-.142 




-.211 


MCATV 


.235 


.066 


.071 




.068 


Constant 




-47.379 


-49.018 




-54.999 


(Other variabl 
not entered) 


es 











The F-ratio for the difference between groups as distinguished 
by the discriminant analysis is significant at the .001 level for the 
difference between patient care and teaching/research. The F-ratios 
for the differences between administration and either of the other 
two practice types are not significant at the .05 level. 

Differences between the coefficients across groups for a given 
variable represent the estimated effect of that variable on the 
relative probabilities of membership in those groups. 

The variables are listed in the order in which they entered the 
discriminant analysis. 

The significance level for each variable given in the second 
column is the statistical significance of that variable's contribution 
to separating the groups. 



the physician's first residency. And this variable (funding for the surgical specialties) 
is also associated with a choice of a teaching or research career. See Tables 11 and 
12. 

With no funding variables included, the discriminant analysis correctly clas- 
sified 46.7 percent of the physicians into the three types of practice groups (patient 
care, administration, and teaching plus research). With funding associated with the 
physician's undergraduate institution added, the discriminant analysis correctly 
classified 49.4 percent; and with funding associated with the institution of the first 
residency added, the discriminant analysis correctly classified 50.4 percent of the 
physicians. We would expect to classify 33.3 percent correctly with random assign- 
ment. And although only about 10 percent of the physicians in the sample are in 
teaching or research, the discriminant analysis correctly classified about 60 percent 
of the physicians in that category. 

If we do the discriminant analysis using only the funding variables without the 
background variables, only 26 percent of the physicians are correctly classified into 
the four specialty groups. And no variables even enter the discriminant analysis in 
the case of predicting type of practice from funding to the undergraduate medical 
institution. 

In sum, federal research funding is to a slight degree related to choices among 
four broad categories of medical specialties, but personal characteristics appear to 



39 



Table 11 



DISCRIMINANT ANALYSIS RESULTS FOR EFFECT ON SPECIALTY CHOICE OF FEDERAL 
RESEARCH FUNDING TO THE INSTITUTION WHERE THE M.D. WAS TAKEN 





Significance 

level d 


Classification Function coefficients 


Variables 


Internal 
Medicine 


Other 
Primary Care 


Surgical 
Specialties 


Other Non- 
primary care 


Sex 


.005 


-14.298 


-13.261 


-14.707 


-14.071 


RANK 


.021 


.029 


.026 


.028 


.027 


FDOPC 


.038 


-151.228 


-144.629 


-152.705 


-151.241 


FDCST 


.001 


8.193 


7.864 


8.365 


8.212 


GPATOTAL 


.009 


42.455 


42.354 


41.146 


42.000 


MCATG 


.028 


.049 


.048 


.045 


.048 


GPASCI 


.127 


-1.624 


-2.114 


-.986 


-1.270 


MCATV 


.093 


.008 


.008 


.011 


.009 


MCATQ 


.133 


.082 


.082 


.081 


.080 


FDSS 


.277 


99.168 


95.521 


100.299 


100.020 


FDONPC 


.156 


-26.870 


-26.445 


-27.747 


-27.539 


SELINDX 


.279 


3.256 


3.238 


3.286 


3.333 


AGE 


.316 


11.662 


11.669 


11.712 


11.624 


Constant 




-276.246 


-272.072 


-273.687 


-273.490 


(Other variables 
not entered) 











The F-ratios for the differences between groups as distinguished 
by the discriminant analysis are all significant at the .025 level. 

Differences between the coefficients across groups for a given 
variable represent the estimated effect of that variable on the 
relative probabilities of membership in those groups. 

The variables are listed in the order in which they entered the 
discriminant analysis. 

The significance level for each variable is the statistical 
significance of that variable's contribution to separating the groups. 



be the principal determinants of these specialty choices. There is no apparent rela- 
tionship between research funding and the choice between primary care and nonpri- 
mary care specialties. Indeed this dichotomy is not a useful one in examining factors 
that influence specialty choices because the characteristics of individuals in differ- 
ent specialties within these broad categories differ almost as much as those in 
specialties between the categories. In the choice of a career in medical training or 
research as in the case of specialty choice, personal characteristics of physicians 
seem to be the principal determinants of a career in academic medicine or research. 



40 



Table 12 



DISCRIMINANT ANALYSIS RESULTS FOR EFFECT ON SPECIALTY CHOICE OF FEDERAL 
RESEARCH FUNDING TO THE INSTITUTION WHERE THE FIRST RESIDENCY WAS TAKEN 



Classification Function Coefficients 



Significance Internal Other Surgical Other Non- 
Variables level Medicine Primary Care Specialties primary Care 



Sex 


.004 


-8.239 


-7.246 


-8.607 


-8.043 


RANK 


.018 


.034 


.032 


.033 


.032 


SELINDX 


.135 


1.567 


1.476 


1.532 


1.558 


GPATOTAL 


.046 


23.864 


24.093 


22.725 


23.865 


GPASCI 


.121 


.931 


.484 


1.572 


1.156 


MCATG 


.113 


.059 


.059 


.056 


.060 


MCATV 


.068 


.035 


.035 


.039 


.036 


FDONPC 


.113 


-.802 


-1.134 


-1.231 


-.917 


FDIM 


.020 


-.552 


-.727 


-.695 


-.824 


FDCST 


.307 


.209 


.356 


.370 


.393 


Constant 




-76.494 


-73.584 


-73.742 


-76.470 


(Other variables 










not entered) 













The F-ratios for the differences between groups as distinguished 
by the discriminant analysis are all significant at the .005 level. 

Differences between the coefficients across groups for a given 
variable represent the estimated effect on the relative probabilities 
of membership in those groups. 

The variables are listed in the order in which they entered the 
discriminant analysis. 

The significance level for each variable is the statistical 
significance of that variable's contribution to separating the groups. 



41 



APPENDIX B 



STATISTICALLY SIGNIFICANT 
UN PATIENT 



Appendix II 
Table B-l 



COEFFICIENTS FROM REGRESSIONS OF HOUSE STAFF SIZE 
LOAD AND NT II FUNDING; SIX SAMPLES 3 



Dependent 






Est tin 


He of: 


Coef f icient 


Estimate for 


Variable 


Year 


littereept 


i 


^ 


OPV 


FUND 


RESIDENTS 


1967-1968 


— 


.00019 
( . 00006 ) 


— 


.00021 
(.00006) 


— 




1968-1969 


8.20 


.00027 





211 


.002 3 






(4.00) 


(.00008) 




(.00006) 


(.0013) 




1969-19 70 


11.35 

(4.03) 


— 


— 


.00026 
(.00007) 


.0030 
(.0014) 




1971-1972 


8.33 


.00054 


-.000035 


.00013 









(4.62) 


(.00007) 


(.000008) 


(.00005) 






1972-1973 


— 


.00046 
(.00007) 


-.000022 
(.000006) 


— 


— 




1973-1974 


13.62 
(5.33) 


— 


— 


.00014 
(.00005) 


— 


INTERNS 


1967-1968 


- 


.00029 
(.00007) 


-- 


- 


- 




1968-1969 


9.36 


.00027 


-.000015 


.00012 


— 






(4.25) 


(.00008) 


(.000008) 


(.00006) 






1969-1970 


9.64 
(4.10) 


.00055 
(.00009) 


-.000039 
(.000009) 


- 


— 




1971-1972 


11.84 
(4.07) 


.00042 
(.00006) 


-.000026 

(.000007) 


— 


— 




1972-1973 


10.10 
(3.33) 


.00033 
(.00004) 


-.000018 
(.000004) 


— 


— 




1973-1974 


5.50 
(3.24) 


.00027 
(.00007) 


— 


— 


— 


TOTAL HOUSE 


1967-1968 


12.00 


.00049 





.00028 





'STAFF 




(7.08) 


(.00012) 




(.00011) 






1968-1969 


17.56 


.00054 


-.000026 


.00032 


— 






(6.93) 


(.00013) 


(.000013) 


(.00010) 






1969-1970 


20.99 

(7.21) 


.00067 
(.00016) 


-.000035 
(.000016) 


— 


— 




1971-1972 


20.19 


.00096 


-.000061 


.00013 









(6.38) 


(.00010) 


(.000010) 


(.00007) 






1972-1973 


14.70 
(7.37) 


.00079 
(.00010) 


-.000040 
(.000008) 


— 


— 




1973-1974 


19.12 


.0004 3 





.00015 









(7.06) 


(.00015) 




(.00007) 





.64 



Standard Errors in parentheses; 
90 percent confidence (using a two-t; 

b Funding in ?1,000. 



all reported coefficients are significantly nonzero with at least 
iled test). 



42 



REGRESSION RESULTS: HOUSE STAFF SIZE ON PATIENT LOADS, 
FACULTY SIZE AND N1H FUNDING; THREE SAMPLES 3 



IVpendi'iit 
Variable 


Year 


Intercept 


Coefficient of 
AA ■ LOG(ALS) 


OPV 


FAC 


FUND b 


R 2 


RESIDENTS 


1971-1972 


8.39 t 


.0015 „ 


.00013 M 





.0017 


.71 






(4.67) 


(.0002) ( 


.00005) 




• (.0022) 






1971-1972 


1.00 
(5.02) 


.0013 . i4 
(.0002) ( 


.00014 SJ 
.00005) 


.188 M 
(.061) 


-.0011 
(.0023) 


.74 




1972-1971 


4.56 


•° 015 ** 


.00012 t 


— 


.0019 


.65 






(5.13) 


(.0002) ( 


.00007) 




(.0014) 






1972-1973 


-1.71 


.0014 ... 4 


.00010 


•I 94 ** 


-.0011 


.68 






(5.36) 


(.0002)" ( 


.00007) 


(.068) 


(.0018) 






1973-1974 


11.74 . v .. 


.0013 M 


.00012 H 





.0005 


.59 






(5.53)'" 


(.0002) " ( 


.00005)' 




(.0014) 






1973-1974 


7.26 


.0012 „ 


.00010 ^ 


.195 iit 


-.0019 


.62 






(5.67) 


(.0002) ( 


.00004) 


(.081) 


(.0016) 




INTERNS 


1971-1972 


11.44 ,. s 
(3.94)"" 


.0012 44 
(.0001) 


- 


- 


-.0008 
(.0018) 


.60 




1971-1972 


5.90 


.0011 M 





.142 4; .. 


-.0028 


.63 






(4.28) 


(.0001) 




(.052) 


(.0019) 






1972-1973 


(3.28) 


.0011 „ 
(.0001) 


— 


— 


.0016 t 
(.0009) 


.60 




1972-1973 


7.02 tt 


.0009 is 


— 


.066 


.0006 


.60 






(3.57) 


(.0001) 




(.045) 


(.0011) 






1973-1974 


5.04 
(3.14) 


.0011 iJt 
( .0001) 


— 


— 


.0009 
(.0008) 


.68 




1973-1974 


2.34 


.0010 M 





•105 ** 


-.0004 


.70 






(3.27) 


(..0001) 




(.047) 


(.0001) 





'Standard errors in parentheses; one asterisk denotes statist 
90 percent confidence; two asterisks denote 95 percent confidence. 



ically nonzero coefficient, 
Two-tailed tests. 



b 



Funding in $1,000. 



III. THE EFFECTS OF FEDERAL BIOMEDICAL RESEARCH 

PROGRAMS ON THE CHARACTERISTICS OF 

DEPARTMENTS 



The medical center activities of education, research, and patient care are to a 
large extent organized along academic departmental lines, although interdiscipli- 
nary research is common. Basic science departments have responsibility for Ph.D. 
programs in their fields and for specific portions of the undergraduate medical 
education curriculum. Clinical departments have responsibility for specific compo- 
nents, "rotations" in the clinical training of undergraduate M.D. students, for the 
training of interns, residents, and fellows in their medical fields, and for the supervi- 
sion of the various inpatient services in major teaching hospitals. 

In this section, we examine the effects of federal biomedical research programs 
on important characteristics of academic departments. First, we examine determi- 
nants of department size, including education, research, and patient care programs. 
Second, we examine evidence of the effects of research programs on faculty involve- 
ment in patient care activity. Third, we examine changes in patterns of support for 
faculty salaries; and, finally, we examine the effects of a department's research 
activity on the salary levels of its faculty members. 



DETERMINANTS OF DEPARTMENT SIZE 

The Model 

The number of full-time faculty has grown with exceptional rapidity since World 
War II. The research responsibility assumed by medical schools — and the availabil- 
ity of federal funding for research — is the most frequently cited explanation for this 
development. To test this hypothesis, we need to develop a model of the determi- 
nants of department size that takes account of research as well as other factors that 
might affect the number of faculty members in a department. 

For a basic science department, such a model might have the following specifica- 
tion: 

F = b + b,S + b 2 R + b 3 G + b 4 SR + 

b 5 SF + b 6 RG + b 7 SRG (1) 

where F = the number of full-time faculty 

b = a constant term 
bi through 

b 7 = the regression coefficients 

S = total medical student enrollment 

R = NIH awards for research 

G = graduate student enrollment 



43 



44 

Equation (1) relates the number of full-time faculty to the number of medical stu- 
dents, the amount of NIH research funding received, and the number of basic 
science students. If, in fact, teaching and research are joint outputs, then the coeffi- 
cients on the interaction terms (b 4 through b 7 ) should be negative and statistically 
significant. 

The model might be modified for a clinical department by substituting house 
staff for graduate students and adding an explanatory variable P to account for the 
patient care load of the clinical faculty. The additional interaction terms need to be 
added to account for joint production. The clinical department model would then 
have the following specification: 

F = b 8 P + b 9 PS + b 10 PR + b„PH + 
b 12 PRS + b 13 PHR + b 14 PHS + b 15 PHSR. (2) 

The coefficients on the interaction terms (b 4 through b 7 ) and (b 9 through bi 5 ) take 
account of the clinical department joint production process that includes patient 
care (P). 

Data and Results 

To estimate the model we use data for FY 1973 1 from the AAMC faculty roster, 
the AMA Directory of Approved Internships and Residencies, the Journal of the 
American Medical Association, and the NIH IMPAC file. Equation (1) is applied 
separately for each major type of basic science department; thus F, R, and G refer 
to faculty, research awards, and graduate students specific to a department. Equa- 
tion (2) is applied to data from the department of medicine using patient load data 
from the internal medicine services of major teaching hospitals. 2 Unfortunately, 
technical problems prevent us from providing estimates of the coefficients on the 
interaction terms. 3 

The results of estimating the truncated equation are presented in Table 13. They 
indicate strong support for the notion that both M.D. enrollment and research 
awards play a statistically significant role in the determination of faculty strength 
in the basic science departments; the role of graduate student enrollment is less 
consistent. 

A better idea of the magnitude of these effects (as opposed to their statistical 
significance) is given in Table 14, which presents elasticity of faculty strength with 
respect to each of the explanatory variables — that is, the percentage change in 
faculty strength that can be expected for each 1 percent change in the explanatory 
variable. In general, a 1 percent change in medical student enrollment has about 
twice the effect of a 1 percent change in research funding. Both elasticities, however, 
are usually much less than one. Thus, funding would have an appreciable but less 
than proportionate effect on faculty strength. 

' This is the latest year for which complete data were available at the time of writing. Complete data 
are now available for FY1974 and will be used in the final report. 

2 We construct a patient load variable that is total annual admissions times the natural log of average 
length of stay. The log of length of stay is used because the physician inputs per day of hospitalization 
are believed to be less the longer the stay. 

1 There is strong multicollinearity between the interaction terms and the other independent variables. 
While the estimated coefficients usually had the right signs, their standard errors were unacceptably 
large. 



45 



Table 13 

CROSS-SECTION REGRESSION ANALYSIS OF DEPARTMENT SIZE 
(Standard errors In parentheses) 



Regression coefficients 



Constant 
e 

Department Number Term 



Anatomy 105 3.65 

Biochemistry 101 3.77 

Microbiology 98 4.35 

Pharnacology 97 5.9 

Physiology 83 4.8 

Medicine 79 15.7 





NIH Research 








Medical 


and Training 


Numbers of 






Student 


Grants to 


Graduate 


Patient 


R 2 


Enrollment 


Departments 


Students 


Load 


.016 a 


9.9 a 


.10 


d 


.60 


(.002) 


(1.8) 


(.06) 






.012 a 


11. 6 a 


.03 


d 


.42 


(.003) 


(2.1) 


(.04) 






.006 a 


4.4 a 


.15 a 


d 


.51 


(.002) 


(1.7) 


(.03) 






.001 


8.3 a 


.23 a 


d 


.54 


(.002) 


(1.3) 


(.05) 






.009 a 


9.8 a 


.io b 


d 


.47 


(.002) 


(2.2) 


(.05) 






.018 


13. 8 a 


.ll a 


22.4 


.37 


(.024) 


(2.81) 


(.063) 


(19.05) 





Significant at 1%. 

Significant at 5%. 

c Total annual admissions times the natural log of average length of stay. 

Applicable only to department of medicine. 

e Data for all departments except medicine are for FY 1973; medicine 
department data are for FY 1972. 



Table 14 
ELASTICITIES, 1974 



Anatomy 

Biochemistry 

Microbiology 

Pharmacology 

Physiology 

Medicine 



.54 


.10 


- 


.40 


.27 


- 


.26 


.08 


.24 


- 


.17 


.23 


.33 


.18 


.10 


_ 


.29 


.21 



46 



We have also tried a modified version of Eq. (2) for medicine departments in Fy 
1972, introducing a variable describing the patient load (annual admissions times 
the log of the average length of stay; for a justification of patient load variables of 
this kind, see the discussion of the determinants of house staff size, above). The 
results are similar to those for basic science departments: NIH. research funding is 
positively and significantly related to faculty size, as is number of house staff, but 
the number of medical students is not. The number of faculty is positively associated 
with the patient load variable, but not in a statistically significant fashion. This may 
reflect the omission of volunteer faculty in our counts of department members. 
Volunteer faculty provide a limited amount of teaching; they also admit patients to 
the teaching hospital. Thus, a school with large numbers of volunteer faculty will 
have a higher value for its patient load variable but no corresponding increase in 
its (observed) faculty strength. 

These results pertain to a cross section of data for FY 1973. Relationships 
observed in the cross section may not, in fact, predict actual changes over time. To 
examine this possibility we ran Eq. (1) using changes in faculty strength in basic 
science departments between FY 1971 and 1974; and changes in the number of 
medical students, amount of research funding (in constant dollars), and number of 
graduate students. The results generally confirm what is presented in Table 13: that 
is, changes in research funding are the most constant predictor of changes in faculty 
strength, while changes in student load are less frequently associated with changes 
in faculty size. 



FACULTY INVOLVEMENT IN CLINICAL CARE 

Objectives and Limitations 

In principle, there are two potential effects of research funding on faculty in- 
volvement in clinical patient care: (DA change in research funding may lead to a 
change in faculty size with consequent effects on both teaching and patient care; and 
(2) a change in funding for research may lead to a change in the allocation of a given 
amount of faculty time among teaching, research, and patient care. The preceding 
subsection examined the first of these potential effects. Here, we consider the effects 
of research funding on patient care when faculty size has already been determined. 
Therefore, to determine the overall effects, results from this and the preceding 
subsection should be considered jointly. 

Availability of data places certain important limitations on this analysis: First, 
we do not have data that reflect actual involvement in patient care by faculty or, 
for that matter, by house staff. Instead, the patient care data measure only the total 
patient loads and average lengths of stay for all hospitals affiliated with each medi- 
cal school in the sample; faculty involvement may be limited to some share of such 
patients, and the share may vary among medical schools, but we cannot test these 
hypotheses. Second the analysis deals only with internal medicine services; behav- 
ior might differ in other hospital services. Third, the variables used in the analysis 
could be brought together only for a single year (academic and fiscal year 1971-72). 



47 



Therefore, the analysis describes cross-sectional behavior rather than responses 
over time, and the cross-sectional results may not be representative of behavior in 
other years. 

Data Sources and Variables 

The variables used in the analysis are: 

(1) Annual admissions to all internal medicine services in hospitals affiliated 
with each medical school (AA); 

(2) Average daily census in all internal medicine services of hospitals affiliated 
with each school (ADC); 

(3) Outpatient visits in internal medicine (OPV); 

(4) Numbers of full-time faculty in the department of medicine in each school 
(FACFT); 

(5) Numbers of part-time faculty in each department of medicine (FACPT); 

(6) Numbers of volunteer faculty in each department of medicine (FACVOL); 

(7) Research funding received by faculty in each department of medicine 
(DFUND); 

(8) Research funding of internal medicine services of affiliated hospitals 
(HFUND); 

(9) Residency positions available, as a proxy for program size (RES); 
(10) Internships available, as a proxy for program size (INT). 

These variables are the same as those used in Section II to analyze effects of 
research funding on house staff size — but with an important distinction: In Section 
II, we assumed that current house staff size in each year is determined in part by 
patient loads in a previous year. Thus, the patient load variables used to analyze 
house staff size in 1971-72 reflect patient loads actually encountered in 1969-70. In 
contrast, here we analyze the relationships between current patient care outputs 
and current faculty and house staff inputs. 

To do this, we had to identify one or more years of data for which current patient 
load, faculty size, funding, and house staff data were available. Faculty data were 
available for 1970-71, 1971-72, 1972-73, and 1973-74; funding data were available for 
1966-67 through 1973-74; house staff data were for 1967-68 through 1973-74, but 
omit 1970-71; patient load data were for 1965-66 through 1971-72, but omit 1968-69. 
The only year for which all of the kinds of data are currently available is 1971-72. 
The means and standard deviations of variables in the file are listed in tables in 
Section II. 

Analysis and Results 

Although not fully specified, the model used in this analysis is a production 
model: We postulate that patient care is an output of a production process using 
faculty and house staff as inputs. Research funding enters the model as follows: We 
postulate that the faculty input to patient care is a function of both faculty size and 
time devoted to research, with research funding used as a proxy for the latter. This 
reasoning leads to a specification of the following general form: 



48 



PATIENT CARE INPUT = flFACULTY, RESEARCH FUNDING, 

RESIDENTS, INTERNS), (1) 

where f denotes a general functional relationship. 

We are interested in whether the coefficients of research funding variables are 
nonzero and, if so, whether they are positive or negative. If positive, the coefficients 
would suggest that, given faculty and house staffsize, research contributes to patient 
care output; and if negative, the coefficients would suggest that research competes 
with patient care as a production activity. 

A number of specific forms of Eq. (1) were used in preliminary analysis. For 
example, we initially used three categories of faculty (FACFT, FACPT, and FAC- 
VOL). However, given the high correlations among the categories within such a 
small sample, attempts to introduce greater detail into the specification produced 
high standard errors for the coefficients and, therefore, little basis on which to 
evaluate the results. In obtaining the results reported below, we use a single mea- 
sure of faculty size (FAC = FACFT + FACPT + .1 VOL) and a single measure of 
funding per faculty member (FUND/FAC). 

The measurement of patient care output leads to some alternative specifications. 
The regression techniques used here allow for only a single dependent variable, yet 
we have three measures of patient care output (admissions, total days of care, and 
outpatient visits). Initial analysis suggested that the number of outpatient visits is 
not highly correlated with either faculty size or funding, although it is highly 
correlated with the remaining measures of patient care output. The results reported 
here either omit outpatient visits or treat it as an explanatory variable; by placing 
outpatient visits among the explanatory variables, we control for the use of some 
share of faculty and house staff inputs in outpatient care, assuming that the provi- 
sion of such care is not affected by research funding or by faculty size. 

Both annual admissions and average length of stay do reveal some correlation 
with faculty size. Since total days of care reflect both admissions and length of stay, 
we initially used total days as the measure of patient care output. However, the 
coefficient of faculty size proved to be negative when days was the dependent varia- 
ble. On balance, the negative correlation between length of stay and faculty size 
appears to outweigh the positive correlation between admissions and faculty size. 

A negative coefficient for an input in a production relation violates common 
sense and suggests specification error. Moreover, the specification of days as the 
dependent variable is inconsistent with the model used in analyzing house staffsize, 
which suggests that fewer inputs are used in providing care on days near the end 
of stay than on days near the beginning of stay. 

A measure of output that places less weight on days at the end of stay can be 
specified by taking the natural log of length of stay. Multiplying this by the total 
number of admissions results in a measure of total patient output that weights 
patient-days more heavily in hospitals with shorter lengths of stay. Replacing days 
with this new variable improved the properties of the regression equation without 
any substantive change in the conclusions regarding effects of research funding. 

Despite the large number of alternative specifications considered, we found little 
evidence of a statistically significant relationship between research funding and 
patient care output for the 1971-72 cross-section sample. The following results are 



49 



generally illustrative of those found in equations whose specifications satisfied in 
general criteria developed above: 4 

AA-LN(ALS) = 3164) + 61.85) FAC + 203.01 RES 
(4040) (41.73) (61.23)** 

FUND 
+ 262.89 INT - .0252 Z,™ (2) 

(84.63)** (.1234) FAC 



R 2 = .60 

The values in parentheses are the standard errors for the corresponding coefficients; 
double asterisks denote significance at 95 percent or better (using two-tailed tests). 
The standard errors indicate that the coefficient of funding (measured in $1,000) is 
not statistically nonzero at any of the usually accepted confidence levels. Moreover, 
the FAC variable also fails to yield a statistically significant coefficient. Therefore, 
there is little basis on which to calculate how patient care might be affected by 
research funding even through its effect on faculty size. 

Limitations 

This analysis was based on a small sample of medical school departments of 
medicine for a single year and uses crude measures of patient care output. For these 
reasons, the results of the analysis are potentially subject to considerable error and 
do not offer a strong foundation for predicting the future implications of changes in 
NIH funding policy. What can be said of the results is that they do not reveal a 
strong relationship between research funding and patient care output, holding 
faculty size constant, for 1971-21. However, the results also reveal little relationship 
between patient care output and faculty size. Thus it appears that the variables, the 
observations, or the empirical methods are inadequate for proper analysis of this 
issue. 



RELIANCE ON NIH FUNDS FOR FACULTY SALARY 
SUPPORT 

Our analysis has shown that NIH funds have strongly influenced the faculty size 
of medical school basic science departments, and that training grants have affected 
enrollment in Ph.D. programs. Thus there is little doubt that the characteristics of 
individual departments have been determined in substantial part by federal bi- 
omedical research and training programs. This indicates that academic medical 
centers have responded to federal program influence by changing their internal 
structure, but it does not provide clear evidence of how vulnerable centers are to 
changes in federal programs. Perhaps the best indicator of departmental vulnerabil- 
ity to changes in federal programs is their reliance on these programs for faculty 

4 The variables whose codes are not denned elsewhere: AA x LN(ALS) = annual admissions times 
the log of average length of stay; FAC = FACFT + FACPT + .1 VOL. 



50 

salary support. The hiring of a faculty member implies a commitment by a center 
of at least several years' duration; in the case of tenured faculty, the commitment 
is of much longer duration. 5 

To the extent that uncertain sources of support from outside organizations, such 
as NIH — so-called soft money — are used to meet a center's firm commitments to 
faculty salaries, the center is vulnerable to changes over which it has little control, 
and it must find other sources of funding to cover its commitments if soft funding 
is cut back. 

We hypothesize that the reliance on soft money for faculty salaries may vary 
within an institution by departments, by faculty rank, and over time. We might also 
expect to find differences between centers depending on their success in obtaining 
research and other soft funding and on whether they are private or state-owned 
institutions. Our analysis examines data on department budgets to determine 
whether consistent patterns exist in the reliance on soft funds for faculty salaries. 

Data Sources 

Analysis of the dependence on soft funds for faculty salaries can be performed 
using both data on sources of support for individual faculty members and aggregate 
data on funding sources for department salaries. The latter are easier to obtain, but 
they may mask differences among departments in their treatment of faculty of 
different rank. Both types of data must be obtained from individual centers because 
there is no central source of such data, and if there were, differences in accounting 
definitions would make them suspect. 

The data used in the analysis reported below were obtained from the ten aca- 
demic medical centers in the earlier Rand study and updated for recent years. 6 In 
the results reported, the data are in most cases drawn from individual faculty 
sources of support, and department data are aggregations of individual salary sup- 
port data. In several cases, however, the sources of funds for faculty salaries were 
not available on an individual basis and total department salary data were provided 
by the school. 

Analysis and Results 

It is possible to obtain a broad picture of sources of support for faculty salaries 
and differences across institutions and departments from simple tabulations from 
medical center budget data. Such tabulations require no statistical analysis. They 
do require extensive examination of medical center financial accounts and the devel- 
opment of a common framework for making cross-school comparisons. Aggregation 
of multiple categories of funds (e.g., NIH research grants, program project grants, 
center grants, career development awards, etc.) into broad classes of funds (e.g., 
HEW research) also facilitates comparisons. 

5 In general, tenure is of less significance in academic medical center departments than in university 
departments, and this is particularly true of clinical departments. Even so, centers consider commitments 
to senior faculty as being of a longer term nature than those to junior faculty. 

6 Data are not available for all ten institutions because the responsibility for faculty compensation 
is in some cases divided among center components — the medical school, teaching hospitals, research 
foundations, departmental group practices partnerships — such that it is impossible to obtain unambigu- 
ous data on total compensation for some individuals. 



51 



The percentages of faculty salaries supported from seven broad classes of funds 
in eight medical schools are presented in Tables 15 and 16 for departments of 
medicine and biochemistry. It is apparent that there is substantial variation across 
schools and between these two types of departments. Such variation exists in the 
data for other departments and institutions. 

It is difficult to observe consistent trends over the time in the data. The only clear 
trend is the increased reliance on patient care revenue for faculty salary support in 
all clinical departments. This same trend is apparent in the aggregate data on 
medical school budgets, and it can be explained by several important interrelated 
changes in recent years. First, Medicare and Medicaid programs have turned many 
of the pre-1965 charity patients of medical centers into "paying" patients, thus 
generating revenue for academic physicians. Second, the full time clinical faculties 
of medical centers have grown substantially in recent years, and part of this growth 
has been more apparent than real, the result of a change in status of physicians who 
in earlier years practiced in academic medical centers but volunteered their teach- 
ing services. Third, the institutional control over practice income has increased in 
most centers during the past eight years. Taken together, these factors account for 
some apparent and some real increase in reliance on practice fees for clinical faculty 
compensation. 

Apart from the increased reliance on practice income, our data show no clear 
pattern among schools, across departments, or over time in the reliance on NIH 



Table 15 







SOURCES 


OF 


SUPPORT 


FOR FACULTY 


SALARIES: 














DEPARTMENT OF MEDICINE 


















FY 


1973-74 




















(Percent) 












Source 


School A 


School 


B 


School 


C School 


D b 


School E 


School F 


School G b 


School H b 


General funds 


36.5 


38.7 




38.3 


7.4 




26.2 


50.5 


7.2 


55.0 


Federal research 


















32.7° 




grants 


25.6 


5.2 




20.5 


29.6 




17.2 


11.8 


2.6 


Federal training 






















grants 


3.0 


4.1 




5.5 


1.1 




7.8 


17.8 


0.0 


0.0 


Other federal funds 


0.0 


0.0 




5.0 


1.2 




0.0 


0.0 


0.0 


0.0 


Foundation grants 


1.7 


1.4 




3.2 


6.0 




20.0 


1.5 


5.2 


0.0 


Patient care 






















revenue 


30.3 


50.5 




26.2 


46.4 




28.8 


17.0 


55.9 


42.4 


Other sources 


2.9 


0.1 




1.1 


8.3 




0.0 


1.4 


0.0 


0.0 



Sources: budget data provided by individual medical schools. 

a General funds include state appropriations, university general funds, endowment, tuition and 
federal and state capitation payments. 

Private medical schools. 

C Includes some federal training grant funds. 



52 



Table 16 







SOURCES OF SUPPORT FOR FACULTY SALARIES: 

DEPARTMENT OF BIOCHEMISTRY 

FY 1973-74 

(Percent) 


■ 






Source 


School A 


School 


B 


School 


C School 


D b 


School 


E 


School F 


School G b 


School H 


General funds 


72.6 


70.7 




81.0 




81.2 




46.0 




60.6 


44.4 


95.2 


Federal research 
grants 


21.1 


24.6 




17.8 




15.7 




13.3 




31.6 


47. 8 C 


2.8 


Federal training 
grants 


0.0 


4.3 




0.0 




0.0 




30.8 




6.1 


0.0 


0.0 


Other federal funds 


2.1 


0.0 




0.0 




0.0 




0.0 




0.0 


0.0 


0.0 


Foundation grants 


0.0 


0.0 




0.9 




3.1 




9.8 




0.0 


6.0 


0.0 


Patient care 
revenue 


4.3 


0.4 




0.0 




0.0 




0.0 




1.5 


1.8 


0.0 


Other sources 


0.0 


0.0 




0.3 




0.0 




0.0 




0.2 


0.0 


0.2 



Sources: budget data provided by individual medical schools. 

General funds include state appropriations, university general funds, endowment, tuition and 
federal and state capitation payments. 

Private medical schools. 

Includes some federal training grant funds. 



research and training funds for faculty salary support. However, institutional poli- 
cies appear to be changing in this area. In general, institutions that have been able 
to retain or increase their NIH research support in the face of tighter federal 
budgets have urged their faculty to apply for more salary support on their grants. 
This trend is most pronounced in the state institutions in our sample. Schools that 
have done less well in recent years competing for NIH research funds have tried to 
shift faculty salaries from this funding source to another in order to reduce their 
vulnerability to federal funding cutbacks in the future. 

We have analyzed data on sources of support for individual faculty salaries to 
determine if there are consistent patterns of support for faculty of different ranks, 
departments, and over time. The statistical technique used was analysis of variance 
(with interactions). The dependent variable in all cases was the percent of faculty 
salary on "soft" funds, where soft funds were denned as federal and nonfederal 
research and research training funds. 7 The categorical explanatory variables were 
faculty rank, academic year, and department type. Departments were classified as 
basic science departments, procedurally oriented departments (all surgical special- 

7 We also analyzed the salary source data using an alternative definition of soft funds that included 
federal capitation and other federal program terms. The results were uninteresting in that we could 
explain less of the variance in the proportion of faculty salaries from soft sources using this broader 
definition than the one restricted to research and research training. 



53 



ties, anesthesiology, radiology, ophthalmology, and otolaryngology), and all other 
clinical departments. 8 

The analysis of variances results are presented for five medical schools in Tables 
17 and 18. Table 17 gives data for salary sources of all faculty members including 
those that receive no money from soft sources. From that table one can estimate the 
average proportion of soft salary funding that faculty members of a particular rank 
in a given school would receive in a particular class of department in a given year. 



Table 17 

ANALYSIS OF VARIANCE 

FACTORS AFFECTING PERCENTAGE OF FACULTY 

SALARIES PROVIDED FROM SOFT SOURCES 3 

(Including all faculty) 





School A 


School B 


School C 


School D 


School E 


Grand Mean 


37.3 


37.3 


8.9 


26.6 


28. A 


Year effects . 
Statistical significance 












NS 


NS 


NS 


.078 


NS 


1966 








-5. A 




1967 








-3.2 




1968 


-A. 5 


3.7 




-0.3 




1969 


1.2 


0.8 




-0.6 




1970 


A. 7 


-1.2 




0.1 




1971 


3.0 


1.8 




2.7 




1972 


-0.2 


-1.6 




3.0 


2.7 


1973 


0.0 


-3.0 


-.2 


-0.A 


-l.A 


1974 


-A. 9 


1.2 


.2 




-0.9 


1975 










0.0 


Department and rank effects 












Statistical significance 


.001 


.001 


.001 


.001 


.001 


Basic Science 












Professor 


-20.7 


-26.0 


15.4 


-5.6 


-11.8 


Associate Professor 


31.0 


-7.1 


1.2 


1.9 


13.2 


Assistant Professor 


5.8 


-18.6 


1A.5 


9.2 


11.9 


Instructor 


15.1 


-9.9 


34.2 


-10.6 


-7.8 


Procedural Specialty 












Professor 


-22.9 


-1A.1 


-5.2 


-15.3 


-18.2 


Associate Professor 


-26.5 


-11.3 


-5. A 


-6.3 


-3.4 


Assistant Professor 


-23.2 


-14. 2 


-7.0 


-1.8 


1.1 


Instructor 


-9. A 


-2.0 


2.5 


-23.8 


-20.9 


Other Specialties 












Professor 


-7.9 


-15.2 


-1.1 


-10.6 


2.1 


Associate Professor 


8.8 


5.4 


-3.2 


3. A 


1.4 


Assistant Professor 


13.2 


16.2 


2.1 


16.6 


7.0 


Instructor 


33.0 


17.2 


-A. 9 


20.6 


72.8 


Proportion of variance 












explained 


.21 


.09 


.12 


.10 


.07 



Soft sources include federal and nonfederal research and research training 
funds. 

The numbers in the columns are deviations from the grand mean, adjusted for 
the other independent variables. They are analogous to the coefficients in a 
multiple regression that uses only categorical data as the explanatory variables. 

Blank spaces in year columns indicate no data available for that year. 

Main effects that are not significant at the .10 level are noted as "NS" ; 
significance levels beyond .001 are listed as .001. 



fl The analysis was performed separately with pathology classified as a basic science department and 
as a procedural specialty. Results were not very sensitive to changes in this classification. 



54 



Table 18 

ANALYSIS OF VARIANCE 

FACTORS AFFECTING PERCENTAGE OF FACULTY 

SALARIES PROVIDED FROM SOFT SOURCES 3 

(Excludes faculty with no soft funds) 





School A 


School B 


School C 


School D 


School E 


Grand mean 


59.9 


75.0 


33.8 


53.2 


47.7 


b c 
Year effects , 

Statistical significance 












NS 


.085 


NS 


NS 


NS 


1966 








-4.4 




1967 








-2.8 




1968 


A. 6 


0.1 




-3.3 




1969 


.6 


-0.7 




-1.9 




1970 


3.0 


1.0 




-1.2 




1971 


0.4 


8.5 




1.3 




1972 


-4.0 


0.2 




3.3 


3.0 


1973 


-0.3 


-2.4 


-1.1 


2.8 


0.2 


1974 


-2.1 


-6.5 


1.1 




-1.0 


1975 










-1.7 


Department and rank effects 












Statistical significance" 


.001 


.001 


.001 


.001 


.001 


Basic Science 












Professor 


-31.5 


-23.1 


0.2 


-26.2 


25.5 


Associate Professor 


14.5 


-7.5 


-12.4 


-21.6 


-1.5 


Assistant Professor 


0.7 


-23.4 


2.8 


-11.5 


-5.8 


Instructor 


17.1 


16.0 


26.9 


27.4 


34.0 


Procedural Specialty 












Professor 


-16.9 


-16.5 


-18.3 


4.3 


10.0 


Associate Professor 


-32.9 


-4.9 


-13.9 


15.9 


16.8 


Assistant Professor 


-14.8 


0.0 


-10.8 


18.3 


23.4 


Instructor 


-2.8 


16.2 


34.8 


24.3 


53.6 


Other Specialties 












Professor 


-5.5 


-15.0 


-8.0 


-14.7 


-1.4 


Associate Professor 


-0.1 


-5.1 


-10.4 


5.7 


5.9 


Assistant Professor 


10.2 


3.1 


20.7 


17.9 


6.6 


Instructor 


24.4 


16.7 


17.9 


32.2 


52.7 


Proportion of variance 












explained 


.21 


.07 


.22 


.27 


.16 



Soft sources include federal and nonfederal research and research training 
funds. 

The numbers in the columns are deviations from the grand mean, adjusted for 

the other independent variables. They are analogous to the coefficients in a 

multiple regression that uses only categorical data as the explanatory variables. 

c 
Blank spaces in year columns indicate no data available for that year. 

Main effects that are not significant at the .10 level are noted as "NS"; 
significance levels beyond .001 are listed as .001. 



For example, an assistant professor in a basic science department in school A in 1974 
would receive an average 38.2 percent of his salary from soft sources (37.3 - 4.9 + 
5.8). Table 18 uses data for only those members who received some salary support 
from soft funds. Thus, from Table 18 we would estimate that given he were to receive 
soft fund salary support, the same assistant professor would be expected to receive 
58.5 percent of his salary from soft funds (59.9 - 2.1 + 0.7). 

In general, the results indicate that the junior faculty are somewhat more 
heavily supported with soft money than senior faculty, and this differs among the 
three classes of departments. Year to year differences are not significantly different 
in most schools but one school (D) shows a slight trend that is marginally significant. 



55 



The most striking thing about the analysis is not the existence of statistically 
significant relationships but that these factors do not account for much of the 
differences in soft money support for individual faculty. In none of the cases have 
we been able to explain more than 27 percent of the total variance (R 2 ) in the 
proportion of individuals' salaries supported by soft funds, and the average R 2 s are 
13 percent for all faculty and 19 percent for those receiving some soft salary support. 

Conclusions 

There is little apparent validity to broad generalizations about the dependence 
of academic medical centers on federal biomedical research funds for faculty salar- 
ies. This dependence varies greatly across institutions, and not surprisingly those 
that have been most consistently successful in competing for research funds have 
tended to rely more heavily on this source of support. Although as a matter of policy, 
some state schools have tried to draw more heavily on soft funds in recent years, the 
data show no strong trend in this direction. 

The only consistent trend in sources of support for faculty salaries is the in- 
creased reliance on practice earnings to pay clinical faculty salaries. However, part 
of this change is more apparent than real, resulting from the reclassification of some 
faculty from volunteer to full time and the increased accountability for practice fees. 
The real increase in revenue from this source due to expansion in the patient care 
functions of academic medical centers and public health insurance programs for the 
aged and needy is probably a one-time phenomenon. It does not represent a readily 
expandable source to replace research funds currently used for faculty salaries. 

Analysis of data on individual faculty salaries does not reveal any consistently 
strong pattern of vulnerability to soft funding cutbacks by rank or department type. 
This does not confirm but is consistent with the hypothesis that individual faculty 
differences — among other things, success in research grants competition — account 
for variation in the dependence on soft funds for salary. 



NIH FUNDING AND FACULTY SALARY LEVELS 

We have seen that the importance of federal research funds as a source of 
support for faculty salaries varies across institutions. This variation can be attri- 
buted in part to differences in the research intensity of the institutions, but it is also 
influenced by other institutional factors — in particular, whether the institution is 
public or private. The variation in dependence on research funds for faculty salaries 
begs questions about the effects of research involvement on salary levels. For exam- 
ple, one might hypothesize that the availability of research funds for faculty salaries 
might lead to inflated salaries in research-intensive institutions. 

The analysis in this section is in two parts. The first is an attempt to describe 
the way in which faculty salaries in a particular department are related to the level 
of NIH funding in that department. The second deals with how these salary levels 
change when NIH funding levels change. 



56 



The Data 

The data for both parts of the analysis were derived from the same sources. 
Salary data came from the AAMC Faculty Salary Surveys for academic years 1973- 
74 and 1974-75. From these surveys we extracted average salaries by rank of strict 
full-time professors in 13 departments of each of the 117 medical schools for which 
the AAMC keeps data. 9 Because of the sensitive nature of the raw data, the analysis 
was actually performed at the computation facilities of the AAMC. Grant data by 
department were from the NIH IMP AC file and included data on all types of NIH 
grants. For convenience in this analysis we grouped all grants into two categories: 
Research grants include individual research grants, research support grants, pro- 
gram-project grants, and clinical research center grants; training grants include 
training grants, fellowship grants, and career development awards. Grant data are 
from fiscal years 1973 and 1974. Missing data for some departments reduced actual 
sample sizes well below the maximum possible size. 

Analysis of Salary Levels 

For this analysis average salary level for each professorial rank was regressed 
on the amount of NIH funding per professor, a measure of the relative cost of living 10 
in the locale of the school, a dummy variable indicating whether the school was 
publicly or privately controlled, dummy variables denoting the region of the country 
(northeast, south, midwest, or west) where the school was located, and dummy 
variables indicating the kind of department (anatomy, biochemistry, etc.). Table 19 
gives the description of the variables. Each equation was estimated twice, once using 
NIH research funding and once using NIH training funding. 11 The complete results 
of these regressions are given in Tables 20 and 21. 

By far the largest part of the explained variance in salaries at all ranks is 
accounted for by the department and region dummies and the cost of living index. 
The coefficients of the region and department dummies can be interpreted as devia- 
tions from the average salary of a medicine department in the western region. In 
almost all cases, however, there is also a significant effect due to NIH funding that 
reveals an interesting pattern. For all ranks of professors a larger amount of funding 
per faculty member is associated with a lower average salary. For department 
chairmen the relationship is reversed. 

A possible explanation of this pattern is that in "prestige" departments — those 
with large grants — professors receive some nonmonetary remuneration in the form 
of career advancement, improved reputation, association with stimulating col- 

9 The departments were anatomy, biochemistry, microbiology, pathology, pharmacology, physiology, 
medicine, obstetrics, pediatrics, psychiatry, radiology, surgery, and orthopedic surgery. 

10 This measure was the Consumer Price Index for the city in which the medical school is located when 
such an index was computed, or the CPI for the nearest city. For schools not located near major cities 
the U.S. urban average CPI was used. It should be noted that these CPI measures give only an imperfect 
comparison among cities, since they are constructed primarily to show difference from one year to the 
next within the same city. They are, however, the best measure of relative cost of living available, and 
that is why we used them. 

" It was not useful to include both types of funding in the same equation because two variables are 
highly collinear. (The correlation coefficient between them is in the neighborhood of 0.8.) By estimating 
two equations, we run the risk of attributing to each some of the effect of the other. It will be seen, 
however, that the estimated coefficients for both are similar and the general picture that emerges is not 
highly sensitive to the specification. Regressions with both variables included were run and the results 
did not vary significantly. 



57 



Table 19 
DESCRIPTION OF VARIABLES 



PUBLIC 



1 for publicly controlled schools; for privately 
controlled schools. 



RSRH 



Amount of NIH research tundinu per professor. 



TECH 



Amount of NIH training funding per professor. 



IRSRH 



Increase in NIH research funding per professor. 



ORSRH 



Decrease in NIH research funding per professor. 



ITECH 



Increase in NIH training funding per professor. 



DTECH 



Decrease in NIH training funding per professor. 



Consumer Price Tndex in area of medical school. 



Change in Consumer Price Index in area of medical school. 



Northeast , 

South, 

Midwest 



Dummy variables for region. 



Anatomy, 

Biochemistry. 

Etc. 



Dummy variables for type of department. 



leagues, etc. and are thus willing to work for a lower monetary income. Chairmen, 
however, act as entrepreneurs in building and managing such "prestige" depart- 
ments and in attracting grants. They are paid according to their success. It is 
interesting to note that associate professors give up more in monetary terms than 
do full professors in order to be associated with departments receiving high levels 
of grants. Assistant professors give up still more. 

A fairly high level of grants is associated with a fairly high level of average 
salary for all ranks of professors. To be consistent with the results reported above, 
this implies a different seniority structure in these departments, with the faculty 
being somewhat more "top-heavy" in departments receiving high NIH funding. 12 

Training grants seem to have a slightly more pronounced effect on salary levels 
than do research grants, but because of the closely collinear nature of research and 
training grants, not much should be made of this difference. 



12 Another study, The Federal Government and Academic Medicine: The Effect of Federal Programs 
on Activities and Output, by Grace M. Carter et al., The Rand Corporation, R-1814-HEW (forthcoming), 
reports that the rapidity of promotion of a medical school faculty is not related to the success of that 
faculty member in obtaining NIH research grants. This study did not address the issue of how the level 
of grants in a particular department influenced promotion. 



58 



Table 20 
RESEARCH GRANTS AND FACULTY SALARIES 





Full 


Associate 


Assistant 




All 


Variables 


Professor 


Professor 


Professor 


Chairman 


Professors 


Public 


-1638 b 


-418 


-909 b 


-1978 b 


1.518 




(375) 


(333) 


(258) 


(543) 


(387) 


RSRH 


-.003 b 


-.012 b 


-.012 b 


.003 b 


.003 a 




(.001) 


(.002) 


(.001) 


(.002) 


(.001) 


CPI 


29 b 


27 b 


26 b 


54 b 


15 a 




(7) 


(6) 


(5.0) 


(10.6) 


(7) 


Northeast 


-316 b 


-2864 b 


-3006 b 


-3115 b 


-3095 b 




(762) 


(657) 


(501) 


(1063) 


(792) 


South 


-3621 b 


-3143 b 


-2469 b 


-1781 b 


-366l" 




(552) 


(482) 


(370) 


(800) 


(576) 


Mi'.Iwest 


-4220 b 


-3070 b 


-1765 b 


-1146 3 


-2698 b 




(515) 


(455) 


(359) 


(771) 


(544) 


Anatomy 


-11628 b 


-11076 b 


-9365 b 


-15401 b 


-10273 b 




(739) 


(651) 


(523) 


(1132) 


(770) 


Biochemistry 


-U051 b 


-10574 b 


-8550 b 


-14456 b 


-9712 b 




(738) 


(651) 


(523) 


(1131) 


(784) 


Microbiology 


-11084 b 


-10347 b 


-8840 b 


-14331 b 


-9341 b 




(732) 


(649) 


(523) 


(1129) 


(780) 


Pathology 


-1250 3 


-1236 3 


-1226 b 


-890 


63 




(757) 


(672) 


(539) 


(1228) 


(807) 


Pharmacology 


-10280 b 


-9944 b 


-8428 b 


-14299 b 


-8912 b 




(755) 


(669) 


(531) 


(1140) 


(806) 


Physiology 


-10357 b 


-9858 b 


-8714 b 


-13829 b 


-9045 b 




(731) 


(643) 


(524) 


(1125) 


(772) 


Obstetrics 


248 


932 


-461 


1215 


1466 




(894) 


(764) 


(577) 


(1352) 


(968) 


Pediatrics 


-1807 a 


-2900 b 


-2509 b 


-3574 b 


-3017 b 




(803) 


(724) 


(565) 


(1289) 


(868) 


Psychiatry 


-3068 b 


-2524 b 


-2830 b 


-781 


-1681 3 




(794) 


(712) 


(559) 


(1303) 


(842) 


Radiology 


4761 b 


492 3 b 


4445 b 


5135 b 


4204 b 




(835) 


(750) 


(580) 


(1327) 


(896) 


Surgery 


4760 b 


2785 b 


2017 b 


5913 b 


34 36 b 




(137) 


(724) 


(569) 


(1380) 


(876) 


Orthopedic Surgery 


- 


- 


- 


-43161 b 
(1211) 


-31415 b 
(825) 


Constant 


328 


-4924 


-10003 


-26895 


11242 


R 2 


.537 b 


.54 3 b 


.542 b 


.661 b 


.693 b 



Significant at the .05 level. 
Significant at the .01 level. 



59 



Table 21 
TRAINING GRANTS AND FACULTY SALARIES 





Full 


Associate 


Assistant 




All 


Variables 


Professor 


Professor 


Professor 


Chairman 


Professors 


Public 


-1668 b 


-383 


-988 b 


-1985 b 


33 




(375) 


(337) 


(262) 


(545) 


(388) 


TECH 


-.016 b 


-.021 b 


-.003 b 


.008 


.ooi a 




(.004) 


(.007) 


(.005) 


(.009) 


(.008) 


CP1 


28 b 


28 b 


27 b 


55 b 


15 a 




(7) 


(6) 


(5) 


(10) 


(7) 


Northeast 


-3110 b 


-2918 b 


-3076 b 


-3139 b 


-3087 b 




(761) 


(664) 


(506) 


(1064) 


(793) 


South 


-3611 b 


-3117 b 


-2436 b 


-1831 3 


-3654 b 




(552) 


(486) 


(374) 


(800) 


(577) 


Midwest 


-4208 b 


-3028 b 


-1760 b 


-1176 


-2696 b 




(515) 


(459) 


(363) 


(771) 


(544) 


Anatomy 


-11638 b 


-10542 b 


-9137 b 


-15543 b 


-10309 b 




(736) 


(648) 


(528) 


(1130) 


(769) 


Biochemistry 


-11076 b 


-10234 b 


-8494 b 


-14543 b 


-9735 b 




(736) 


(654) 


(531) 


(1131) 


(783) 


Microbiology 


-11085 b 


-9917 b 


-8669 b 


-14437 b 


-9395 b 




(729) 


(649) 


(529) 


(1127) 


(778) 


Pathology 


-1245 a 


-740 


-1112 3 


-1025 


-36 




(754) 


(670) 


(545) 


(1225) 


(804) 


Pharmacology 


-10218 b 


-9497 b 


-8197 b 


-14407 b 


-8988 b 




(749) 


(669) 


(535) 


(1139) 


(803) 


Physiology 


-10345 b 


-9498 b 


-8521 b 


-13940 b 


-9097 b 




(727) 


(645) 


(529) 


(1123) 


(770) 


Obstetrics 


220 


1425 a 


-258 


1064 


1440 




(892) 


(766) 


(583) 


(1350) 


(968) 


Pediatrics 


-184 5 a 


-2392 b 


-2 326 b 


-3705 b 


-3055 b 




(803) 


(724) 


(571) 


(1287) 


(867) 


Psychiatry 


-3083 b 


-1941 b 


-2579 b 


-942 


-1709 3 




(791) 


(710) 


(565) 


(1301) 


(842) 


Radiology 


4789 b 


5494 b 


4617 b 


4983 b 


4121 b 




(830) 


(748) 


(585) 


(1324) 


(893) 


Surgery 


4788 b 


2851 b 


1873 b 


5781 b 


3361 b 




(812) 


(730) 


(576) 


(1378) 


(874) 


Orthopedic Surgery 


- 


- 


- 


-43330 b 
(1208) 


-31448 b 
(824) 



Constant 
2 



R 



1688 



.538 



-6935 



.536 



-10965 



.533 



-27538 



.660 



10974 



.693 



Significant at the .05 level. 
Significant at the .01 level. 



Analysis of Changes in Salary Levels 

Although the preceding analysis of salary levels gives some indication of how 
faculty salaries and NIH funding are related at present, it tells us nothing directly 
about what would happen to salaries if NIH funding levels were to change. To 
explore this question, we regressed changes in average salaries from one year to the 
next on changes in NIH funding, the level of NIH funding, changes in the cost of 
living, dummy variables to indicate region of the country, and a dummy variable to 
indicate public or private control. As in the analysis of salary levels, all funding data 
were computed on a per faculty member basis. 

Because only two years of salary data were available, we were able to compute 
only one observation for each department. This limited sample size requires that 
these results be viewed as quite tentative. Some patterns do emerge. As always, data 
on changes are subject to wide variation caused by factors not included in our model. 
We have not, therefore, been able to explain very much of the variance in changes 
in salaries. 

We found in preliminary analyses that patterns of change in salaries differ from 
clinical departments to basic science departments. Complete results of the regres- 
sions for clinical departments can be found in Table 22 and those for basic science 
departments in Table 23. In clinical departments there appears to be a positive 
relation between increases in research grants and increases in full professors' salar- 
ies. This relationship does not seem to hold, however, for associate or assistant 
professors. Assistant professors appear to be highly vulnerable to decreases in train- 
ing grants. For each dollar reduction in the amount of training grants per faculty 
member, increases in assistant professors' salaries are reduced 26 cents. In depart- 
ments with high levels of research grants per faculty member, full professors' salar- 
ies have risen less and associate professors' more than in departments with lower 
grant levels. In departments with high levels of training grants, salaries of both full 
and assistant professors have risen more than those in departments with lower 
levels of training grants. Levels of neither type of grant appear to affect the size of 
changes in assistant professors' salaries. 

Very few coefficients fitted for basic science departments are statistically differ- 
ent from zero. Perhaps with more years of data some patterns would emerge, but 
on the basis of what was available for this study there is simply no interpretable 
pattern for basic science departments. 



61 



Table 22 
CHANGES IN FACULTY SALARIES IN CLINICAL DEPARTMENTS 





Full 


Associate 


Assistant 




Variables 


Professors 


Professors 


Professors 


Chairman 


Public 


1380 a 


-1082 


-206 


-2846 




(701) 


(745) 


(620) 


(2737) 


RSRH 


-.010 b 


-.010 


-.007 


-016 




(.003) 


(.011) 


(.005) 


(065) 


TECH 


.089 b 


.069 


.017 


-.227 




(.028) 


(.045) 


(.031) 


(.313) 


IRSRH 


.023 a 


-.018 


.009 


-.116 




(.011) 


(.018) 


(.013) 


(.141) 


DRSRH 


.115 a 


.012 


.015 


.197 




(.058) 


(.014) 


(.011) 


(.294) 


ITECH 


-.175 b 


-.007 


-.011 


.140 




(.051) 


(.083) 


(.058) 


(.579) 


DTECH 


.144 


.069 


.268 a 


.340 




(.158) 


(.134) 


(.135) 


(.567) 


ACPI 


64 a 


17 


37 


166 




(38) 


(38) 


(33) 


(141) 


Northeast 


1230 


-786 


721 


-6504 




(1105) 


(1132) 


(953) 


(4152) 


South 


-1147 


-1189 


-715 


-4044 




(897) 


(924) 


(771) 


(3315) 


Midwest 


-573 


-853 


-1002 


-4156 




(889) 


(901) 


(802) 


(3314) 


Constant 


-7206 


1290 


-3120 


-8264 


R 2 


.238 b 


.059 


.080 a 


.051 


Significant at the 


.05 level. 






Sign: 


ficant at the 


.01 level. 







62 



Table 23 
CHANGES IN FACULTY SALARIES IN BASIC SCIENCE DEPARTMENTS 





Full 


Associate 


Assistant 




Variables 


Professors 


Professors 


Professors 


Chairman 


Public 


-11.9 


487 a 


385 


531 a 




(290) 


(269) 


(250) 


(267) 


RSRH 


.003 


.on a 


-.oio a 


.005 




(.005) 


(.005) 


(.004) 


(.005) 


TECH 


-.031° 


-.000 


.026 a 


.013 




(.016) 


.001 


(.015) 


(.021) 


IRSRH 


-.008 


-.010 


.008 


-.006 




(.012) 


(.008) 


(.005) 


(.006) 


DRSRH 


.000 


-.025 


-.029 


.032 




(.023) 


.021 


(.019) 


(.022) 


ITECH 


.028 


-.026 


-.029 


-.058 a 




(.026) 


.028 


(.023) 


(.028) 


DTECH 


.016 


-.004 


.078 


.008 




(.056) 


(.054) 


(.069) 


(.055) 


ACPI 


9.3 


-3.1 


10.1 


13.1 




(16.1) 


(15.2) 


(13.5) 


(14.9) 


Northeast 


659 


774 a 


451 


1145 b 




(469) 


(435) 


(386) 


(414) 


South 


95.6 


-136 


-370 


243 




(411) 


(398) 


(347) 


(384) 


Midwest 


240 


556 


-197 


610 




(383) 


(382) 


(341) 


(371) 


Constant 


494 


1333 


-493 


-105 


R 2 


.032 


.045 


.042 


.068 a 



Significant at the .05 level. 
Significant at the .01 level. 



IV. THE EFFECTS OF FEDERAL BIOMEDICAL RESEARCH 
PROGRAMS ON INSTITUTIONAL FUNDING 



Federal biomedical research funds are an important source of revenue for aca- 
demic medical centers. However, the importance of biomedical research funding has 
declined relative to other sources. In 1974, federal research funds accounted for 
approximately 21 percent of the total of operating budgets of all centers; ten years 
earlier, the comparable figure was 36 percent. Obviously, the financial structure of 
academic medical centers must have changed substantially in this period, but highly 
aggregated data can provide very few insights into the effects of funding changes on 
particular institutions. 

This section analyzes the effects of funding changes on academic medical centers 
at two levels. The first is the total operating budget of a center exclusive of teaching 
hospitals — the effects of changes in NIH funding on funding from other sources (e.g., 
state appropriations, foundations) for the center as a whole. The second level is the 
internal budget process within a particular center — the effects of changes in funding 
from outside the center (e.g., NIH research, patient fees) on the allocation of funds 
that are under the discretionary control of the dean or vice-president. 



ALTERNATIVE SOURCES OF REVENUE FOR MEDICAL 
SCHOOLS 

The Model 

Recently questions have arisen about what effect the receipt of NIH funding has 
on the efforts and ability of medical schools to find and exploit other sources of 
revenue. For example, are losses in federal funding likely to be made up by increased 
revenues from state governments and foundations? Are losses in federal training 
funds passed on to students in the form of higher tuition? Do other sources of 
revenue view the receipt of NIH awards as an indication of merit and make their 
own awards accordingly, thus multiplying the effect of NIH awards? These are the 
types of questions addressed in this section. 

Inasmuch as this is an exploratory analysis, the conceptual model used here is 
quite simple. The analysis is both cross-sectional and time series, in that we have 
followed the pattern of revenues from nonfederal sources for all medical schools over 
a seven-year period. The principal aim of our analysis is to determine by what 
amount revenues from these sources rose or fell as a result of changes in levels of 
various types of NIH funding. 

For each of 15 alternative sources of funding a regression equation was fitted. 
All equations had the same form, regressing yearly changes in revenues from the 
source in question on several types of variables. Of principal interest are the coeffi- 
cients associated with changes in NIH funding in the current and previous years. 
These coefficients are a measure of the extent to which funds from the alternative 
sources matched or replaced NIH funds over the time period covered by our sample. 
For example, a coefficient of 140 indicates that a $1000 increase in NIH funds 

63 



64 



brought with it an additional $140 of revenue from the alternative source in ques- 
tion, all other things being equal. A negative coefficient indicates that NIH funding 
replaces or is replaced by funds from the alternative source. 

It is possible that medical schools may respond differently to decreases in NIH 
funding than they do to increases due, perhaps, to long-term commitments of the 
schools or other institutional rigidities that make it difficult for schools to respond 
to some stimuli. To allow for this type of behavior, changes in NIH funding have 
been decomposed into two variables, one representing increases in such funding and 
the other decreases. Only one of these variables can be nonzero in a particular 
observation. It is possible that the size of a medical school might affect its behavior 
in seeking money from nonfederal sources. Larger schools may be better able to 
absorb losses or may have more extensive experience in raising funds from a wide 
variety of sources. In an effort to control for these effects, two measures of the 
financial size of a school have been included in the regression, total revenue and 
total NIH grants of a given type. 

Growth in the size of the undergraduate student body and of the faculty should 
also be taken into account. To the extent that increased revenues from tuition may 
reflect only the growth of the student body or rising revenues from the professional 
activities of the faculty may reflect faculty growth, these changes are not of concern 
in this section of the study. We have therefore included in the regressions terms for 
changes in number of students and faculty. Like NIH grants, these have been 
divided into increases and decreases. Finally, dummy variables have been included 
to reflect the differing financial environments faced by public and private schools 
and by accredited and provisionally accredited institutions. 

The equations estimated have the following form.(AR is the change in revenue 
from a particular source. For a discussion of the sources considered, see below.) 

AR = a, + a 2 ACRDT + a 3 NIH + aJNIH + a 5 DNIH + aJNIHL 

+ a 7 DNIHL + a 8 IUG + a 9 DUG + a I0 IFAC + a u DFAC 

+ a 12 TREV + a 13 PUBLIC. 

A description of each of the variables used in this equation is found in Table 24. 

The Data 

All data except NIH grant data are from surveys taken by the Liaison Commit- 
tee on Medical Education (LCME) of the Council on Medical Education, American 
Medical Association and the Association of American Medical Colleges. Most of the 
data for this analysis were drawn from the Institutional Profile System (IPS) file. The 
results of the LCME questionnaire were available for the seven academic years 
1967-68 through 1973-74, allowing nine sources of revenue to be followed through 
this period. For the last three of these years, a richer set of responses was available, 
allowing an additional six sources to be considered for this shorter period. Although 
all data in the IPS are kept for the 117 medical schools in operation during the 
1974-75 academic year, some schools were not in operation during the early portion 
of the period covered by the study and there are often data missing from the survey 
responses. Because the regression equations fitted involve differences and lagged 
differences, it is possible to produce a maximum of five observations per school out 
of seven years of data. The missing data reduce the actual sample sizes to considera- 
bly less than the 585 theoretically possible. 



65 

Table 24 
DESCRIPTIONS OF VARIABLES 



ACRDT 



NIH 



INIH 



DNIH 



INIHL 



DNIHL 



IUG 



DUG 



IFAC 



DFAC 



TREV 



PUBLIC 



Dummy variable; 1 if school is accredited, 
if provisionally accredited. 

Amount of NIH grants. (Various formulations of 
this variable were used. See discussion of data 
below. ) 

Increase in NIH grants from one year to next, if any. 

Decrease in NIH grants, if any. 

Increase in NIH grants, lagged one year. 

Decrease in NIH grants, lagged one year. 

Increase in number of undergraduates, if any. 

Decrease in number of undergraduates, if any. 

Increase in faculty size, if any (full time). 

Decrease in faculty size, if any (full time). 

Total revenue. 

Dummy variable; 1 if school is public or state 
related, if private. 



Data on NIH grants were of three types. The IPS contains data on NIH funding 
for program projects and center grants, and for research grants received by schools 
during a given academic year. These data are available for academic 'years 1967-68 
through 1973-74. The amounts of grants awarded by NIH to institutions during a 
given fiscal year are available from the NIH IMPAC file for FY 1967 through FY 
1973. This file contains data on program project, center, and research grants, and 
on training grants. Also in this file are data on grants made to hospitals associated 
with medical schools. Each regression equation was fitted three times: once using 
AAMC research grant data, once using IMPAC file research grant data, and once 
using IMPAC file training grant data. Since the IMPAC file data are for grants 
awarded in a given fiscal year, the effects of these grants will most likely show up 
most strongly in the academic year immediately following when most of this money 
is spent. Thus, an effect that appears with a lag with the AAMC data should appear 
to be a current effect with the IMPAC data. (This, in fact, is what is observed.) 



66 



Analysis 

Of the 15 sources of revenue considered in this study, six appear to show patterns 
of response (or nonresponse) to changes in NIH funding of sufficient interest to be 
reported in detail here. Each of these will be dealt with briefly. The complete results 
of the regressions are given in Tables 25-31. The remaining nine 1 seem to exhibit 
no relationship to NIH funding. Great care must be taken when attributing changes 
in revenues from nonfederal sources to a causal relation with NIH funding. What 
is revealed by this analysis is only that during the time period of our study revenues 
from some appeared to change with some relation to changes in NIH funding. The 
inference of a causal relationship must rest on other evidence. In some of the cases 
cited below, the relations that appear are probably spurious. In many cases, the most 
that can be said is that the data either are or are not consistent with a causal relation 
among funding sources. 

It should also be stressed that the results reported here are based on a first 
analysis of the data. In a few cases, which will be noted, we see apparently anoma- 
lous results. These are probably statistical artifacts and could perhaps be explained 
by a more extensive examination than was possible in the present study. In any case, 
they provide useful cautionary tales about interpreting these results too finely. 
Despite these blemishes, the picture that emerges is broadly interpretable. 

State Sponsored Research. The level of state sponsored research appears to 
respond to changes in NIH funding in the ways one might expect. Both the AAMC 
data and the IMPAC data on research grants show a negative relation between 
increases in NIH funding and state sponsored research. The AAMC data suggest 
that an increase of $1000 in NIH research replaces about $149 of state research 
money. The figure implied by the IMPAC data is $175. Decreases in NIH funding 
seem to have no noticeable effect. As one might expect, there is no effect due to 
changes in NIH training grants. 

The situation depicted by these results seems to be one in which large amounts 
of NIH money replace small amounts of state money, but not vice versa. Perhaps 
state research funds are used as "seed money" to get particular programs started. 
This possibility should be explored in more depth. 2 

Nongovernment Sponsored Research. Research sponsored by nongovernmental 
sources shows behavior similar to that of total gift revenue. In the short run there 
appears to be a negative relation between this type of research support and NIH 
research funding, but in the longer term money from these sources follows NIH 
money. With AAMC data, an increase in current NIH research of $1000 funding 
brings about a drop of $57 in nongovernment research money. A lagged increase in 
NIH funding of the same size brings a gain of $124. Once again, as in many of the 
cases above, decreases in NIH funding seem not to have an effect. 

This pattern is consistent with the IMPAC data. The early negative effect is not 
observed because when using this data set we observe funding only after a lag. The 

1 These are public and state related appropriations, alumni gifts, business and industry gifts, private 
school subsidies from state and local governments, revenues from intrastate and interstate compacts, 
city-county government appropriations, foundation gifts, state and local multipurpose funds, and nongov- 
ernment multipurpose funds. 

2 This equation is the only one in which there seems to be evidence of serially correlated residuals 
produced by the regression. The serial correlation appears to be positive, and thus it is likely that the 
errors in this equation have been understated and the significance of some effects overstated. Unfortu- 
nately, we were not able to reestimate this equation to eliminate this bias. 



67 



Table 25 
STATE SPONSORED RESEARCH 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


-.803 


15. 6 a 


.768 




(11.6) 


(9.3) 


(26.9) 


INIH 


13.1 


-175.0 b 


-32.1 




(26.9) 


(36.2) 


(216.0) 


DNIH 


-69.3 


-69.0 


-189.8 




(104.2) 


(71.5) 


(135.2) 


INIHL 


-149. l b 


-22.0 


16.6 




(37.8) 


(34.7) 


(169.3) 


DNIHL 


-93.1 


-21.2 


244.1 




(91.7) 


(85.3) 


(251.7) 


IUG 


-665.8 


-395.6 


-507.9 




(854.1) 


(808.5) 


845.9 


DUG 


-1122.7 


-887.3 


-1293.3 




(2275.8) 


(2227.1) 


2327.4 


IFAC 


-1558. 2 b 


-1395. b 


-1495. 2 b 




(436.8) 


(422.1) 


438.5 


DFAC 


1382. 3 a 


1294.4 


1359. 4 a 




(649.0) 


(658.7) 


659.7 


TREV 


5.0 a 


.002 


7.0 




(2.9) 


(.002) 


(2.3) 


ACRDT 


-100985.0 


-22561.9 


17292.2 




(320122.3) 


(180312.5) 


190042.4 


PUBLIC 


21369.1 


28654.6 


19969.1 




(43571.1) 


(40625.6) 


(42700.6) 


CONSTANT 


105995.9 


12131.2 


20154.5 



.160 



.174 



.084° 



Significant at the .05 level. 
""Significant at the .01 level. 



68 

Table 26 
NONGOVERNMENT SPONSORED RESEARCH 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


14.3 


5.8 


3.5 




(11.6) 


(9.4) 


(25.5) 


INIH 


-57. a 


98. 9 b 


134.1 




(28.4) 


(35.8) 


(199.8) 


DNIH 


94.2 


16.4 


38.1 




(109.5) 


(73.9) 


(133.9) 


INIHL 


124. l b 


89. l b 


132.9 




(38.4) 


(35.2) 


(164.0) 


DNIHL 


-4.9 


-81.7 


74.3 




(94.4) 


(78.2) 


(235.8) 


IUG 


391.1 


269.0 


356.3 




(726.6) 


(691.0) 


708.1 


DUG 


1660.7 


1727.3 


1878.5 




(2059.6) 


(2003.9) 


(2055.3) 


IFAC 


-757. 8 a 


-947. 3 a 


-675.8 




(441.8) 


(427.1) 


(436.1) 


DFAC 


-362.5 


21.5 


-491.8 




(705.2) 


(707.0) 


(702.8) 


TREV 


.002 


.001 


.006 b 




.002 


.002 


(.002) 


ACRDT 


-41314.7 


-10292.8 


-26620.9 




(370326.0) 


(162970.7) 


(167735.1) 


PUBLIC 


40818.2 


47566.9 


43485.5 




(38153.2) 


(36234.3) 


(37244.4) 


CONSTANT 


-4428.7 


-37391.6 


-43017.33 


R 2 


.112 b 


.115 b 


.065 b 



Significant at the .05 level. 
Significant at the .01 level. 



69 



Table 27 
NONFEDERAL TEACHING AND TRAINING FUNDS 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


22.3 


24.6 


83. 9 a 




(23.6) 


(19.6) 


(49.3) 


INIH 


.767 


18.4 


-565.2 




(55.9) 


(82.6) 


(380.4) 


DNIH 


10.7 


238. a 


-772. 8 b 




(209.7) 


(139.0) 


(8506.7) 


INIHL 


13.7 


-47.5 


-349.2 




(81.7) 


(71.4) 


(1371.4) 


DNIHL 


254.9 


37.6 


-396.6 




(173.7) 


(147.0) 


(452.2) 


IUG 


2687. l a 


2686. 3 a 


2694. 3 a 




(1375.2) 


(1319.6) 


(1292.6) 


DUG 


-691.2 


-1398.1 


-616.6 




(3596.7) 


(3509.0) 


(3433.1) 


IFAC 


704.5 


683.6 


618.9 




(1080.2) 


(1053.3) 


(980.8) 


DFAC 


97.2 


136.5 


581.6 




(1445.7) 


(1462.6) 


(1358.4) 


TREV 


-.000 


.000 


-.002 




(.005) 


(.005) 


(.004) 


ACRDT 




-46237.2 


-55482.9 




(7.2) 


(441765.3) 


(433679.8) 


PUBLIC 


29427.0 


37616.0 


49564.4 




(78713.6) 


(74672.6) 


(73322.8) 


CONSTANT 


-85603.5 


-42265.1 


-64943.2 


R 2 


.042 


.047 


.083 b 



"Significant at the .05 level. 
Significant at the .01 level. 



70 

Table 28 
TUITION (PRIVATE SCHOOLS) 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


-8.0 


-4.5 


17.5 




(8.1) 


(5.7) 


(16.0) 


INIH 


3.5 


50. l a 


-513. 8 b 




(15.4) 


(20.2) 


(140.2) 


DNIH 


235. 9 b 


1.06 


121.8 




(78.0) 


(47.4) 


(77.0) 


IMIHL 


66. 3 b 


-24.0 


-69.2 




(21.3) 


(20.3) 


(99.9) 


DNIHL 


-.919 


48.7 


127.9 




(54.3) 


(55.2) 


(148.8) 


IUG 


1925. b 


1913. 7 b 


2087. 7 b 




(695.5) 


(687.3) 


(679.5) 


DUG 


2142.8 


1652.3 


2128.1 




(4065.6) 


(4095.9) 


(4035.4) 


IF AC 


-282.5 


-210.2 


-121.6 




(228.2) 


(232.2) 


(228.2) 


DFAC 


-1385. 9 b 


-1614. 4 b 


-1296. 8 b 




(418.9) 


(475.3) 


423.4 


TREV 


.001 


1.46 


1.8 




(.001) 


(1.47) 


(1.3) 



ACRDT 

PUBLIC 

CONSTANT 
„2 



131585.9 



114703.1 



11558.9 



.21310 



.171 



.194 



Significant at the .05 level. 
Significant at the .01 level. 



71 

Table 29 
TUITION (PUBLIC SCHOOLS) 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


-3.7 


5.8 


3.23 




(5.5) 


(4.8) 


(12.4) 


INIH 


-1.1 


4.9 


20.2 




(17.8) 


(20.3) 


(86.8) 


DNIH 


46.7 


28.8 


65.3 




(56.8) 


(39.2) 


(69.8) 


INIHL 


5.8 


-98. 9 b 


-156. 7 a 




(21.9) 


(22.5) 


(82.1) 


DNIHL 


128. 4 a 


3.7 


-312. b 




(58.5) 


(38.2) 


■ (109.6) 


IUG 


479.7 


534. 6 a 


426.9 




(293.0) 


(273.3) 


273.6 


DUG 


-501.0 


-1122.6 


-803.8 




(724.5) 


(695.4) 


(693.8) 


IFAC 


371.0 


-13.9 


106.6 




(538.8) 


(301.0) 


(298.8) 


DFAC 


-267.0 


-46.4 


-61.4 




(354.6) 


(335.3) 


(333.0) 


TREV 


.004 b 


4.5 b 


.003 b 




(.001) 


(1.2) 


(.001) 


ACRDT 


16825.7 


-51856.5 


-55606.2 




(124367.5) 


(68593.0) 


(69012.7) 



PUBLIC 

CONSTANT 
„2 



-15032.5 



37987.0 



42658.5 



.131 



.193 



.183 



Significant at the .05 level. 



b„. 



Significant at the .01 level. 



72 

Table 30 
PROFESSIONAL FEES 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMPAC Data) 


(IMPAC Data) 


NIH 


-10.8 


7.8 


6.9 




(26.5) 


(22.5) 


(57.9) 


INIH 


64.8 


-129.1 


-190.1 




(79.9) 


(86.6) 


(432.6) 


DNIH 


-230.7 


339. 2 a 


-148.2 




(264.3) 


(165.0) 


(304.3) 


INIHL 


-75.5 


.073 


-1.3 




(89.2) 


(79.1) 


(349.4) 


DNIHL 


339.4 


49.7 


-105.1 




(206.9) 


(171.6) 


(531.6) 


IUG 


-2177.8 


-1773.7 


-1737.5 




(2041.0) 


(1986.3) 


(1990.0) 


DUG 


10901. 4 b 


10428. 4 a 


10744. 5 a 




(4169.1) 


(4127.4) 


(4140.8) 


IFAC 


645.6 


882.2 


300.8 




(1228.7) 


(1205.7) 


(1168.0) 


DFAC 


2162.5 


1863.5 


2314.5 




(1664.1) 


(1698.4) 


(1648.5) 


TREV 


.010 


10. 6 a 


7.9 




(.007) 


(6.3) 


(5.5) 


ACRDT 


158335.7 


166663.6 


147443.0 




(742291.3) 


(731234.9) 


(736562.6) 


PUBLIC 


8982.3 


46795.4 


40642.6 




(93496.3) 


(89647.1) 


(90025.4) 


CONSTANT 


130158.2 


80596.4 


91120.0 


R 2 


.066 


.069 a 


.056 



Significant at the .05 level. 
Significant at the .01 level. 



73 



Table 31 
TOTAL GIFT REVENUE 





Research 


Research 


Training 


Variable 


(IPS Data) 


(IMP AC Data) 


(IMP AC Data) 


NIH 


-3.0 


-12.8 


-41.8 




(14.7) 


(10.9) 


(29.8) 


INIH 


4.6 


-115. 8 b 


403. 8 a 




(32.8) 


(43.2) 


(227.3) 


DNIH 


110.2 


-27.7 


-315. 7 a 




(133.9) 


(82.2) 


(166.9) 


INIHL 


-155. b 


143. 7 b 


-290.7 




(47.1) 


(40.9) 


(185.1) 


DNIHL 


-23.6 


39.7 


196.0 




(106.2) 


(96.9) 


(276.2) 


IUG 


-10.2 


-396.2 


-148.5 




(972.2) 


(923.2) 


(944.4) 


DUG 


741.7 


2093.2 


928.7 




(2825.1 


(2726.8) 


(2782.4) 


IFAC 


607.0 


586.9 


504.7 




(506.7) 


(488.0) 


(486.1) 


DFAC 


-4335. l b 


-3964. 8 b 


-4014. 6 b 




(864.5) 


(887.9) 


(853.4) 


TREV 


7.2 a 


.005 a 


.005 a 




(3.4) 


(.002) 


(.002) 


ACRDT 


_ 


-58726.6 


-61549.5 






(194092.7) 


(199616.0) 


PUBLIC 


30371.8 


29007.2 


31975.7 




(46970.8) 


(44167.6) 


(45479.6) 


CONSTANT 


-103259.2 


-32221.98 


-37993.3 


R 2 


.151 b 


.177 b 


.131 b 



Significant at the .05 level. 
Significant at the .01 level. 



74 



positive effects show up as they should, associated with increases in NIH funding. 
Current (that is, for the just ended fiscal year) NIH funding increases of $1000 bring 
an additional $98 in nongovernment research money. The figure for a similar lagged 
increase is $89. 

Changes in training grants have no discernible effect. 

The picture presented here appears to be one of NIH funding in the previous 
year being regarded as an indication of merit by other research funding organiza- 
tions as they make their own grant decisions. There is some evidence that to a small 
degree these sources also make up for short-run decreases in NIH funding. 

Nonfederal Teaching and Training Funds. As might be expected, changes in 
NIH research funding seem to have little effect on the availability of nonfederal 
training funds. The situation is very different, however, when we consider the effect 
of NIH training grants. An increase of $1000 in NIH training grants replaces about 
$596 of funds from other sources in the current year. A loss of $1000 is replaced by 
about $772 from other sources. These same patterns (of a magnitude of about $400 
per $1000 of NIH funds) appear with respect to lagged NIH funding, although the 
variance of these latter estimates is very high and they are in fact not statistically 
significant. 

The pattern that emerges is clearly one of major substitution of NIH funds for 
other training funds and vice versa. 

Total Revenues from Student Tuition 

Tuition revenue is the only source of funding studied that shows markedly 
different behavior for public and private schools. Private schools increased their 
total tuition revenues by about $100,000 more per year than did public schools and 
the patterns of increases are markedly different. 

Changes in NIH research funding appear to be positively related to changes in 
total tuition revenue. Rises in NIH funding seem to have brought with them small 
increases in tuition revenues, and conversely for a fall in NIH funding. When we use 
AAMC data, public schools decreased their total tuition by $128 for each $1000 lost 
in NIH research grants the previous year. Private schools decreased their tuition 
revenues by $235 for each $1000 in NIH money lost during that year and added $66 
per $1000 gained the previous year. When we use IMP AC research grant data, these 
relations are somewhat less marked. An increase of $1000 in NIH funding brought 
a $50 increase in tuition revenue for private schools in the same year and a $99 
increase for public schools the following year. 

When we use IMPAC training grant data, negative effects appear, but curiously 
only for public schools. An increase in NIH funds of $1000 reduces total tuition by 
$156 after a delay of a year and a similar decrease raises tuition revenues by $312. 
For private schools, the relation has been positive with a $1000 decrease in NIH 
training funds being associated with a decrease in tuition of $128 in the same year. 

The picture that emerges, then, is cloudy. There seems to be a small positive 
effect associated with NIH research funding, with public schools taking longer to 
respond than private schools. We have no explanation for these apparent patterns 
of behavior. The results associated with training grants are difficult to understand 
and should probably be attributed to anomalies in the time period in question until 
further study permits a more satisfactory explanation. 



75 

Revenue from Professional Fees. There is a positive relationship between reve- 
nues from the professional services of the faculty of a medical school and lagged 
losses in NIH research funding. Results are similar with both types of NIH data: A 
loss of $1000 in NIH research funding is followed by a loss of $339 in revenue from 
professional services. It is curious that in spite of such a large effect of decreases in 
NIH funding there seems to be no effect due to increases in such funding. Changes 
in training grants seem to have no effect. 

It is difficult to know what to make of these results. Historically, revenues from 
professional services have been underreported by medical schools. The trend, how- 
ever, has been for these reports to become more accurate in recent years. It is 
possible, then, that gains in revenues from this source are randomly reported as 
better accounting procedures are adopted. Losses, a much smaller subset of the 
sample than gains, might be reported quite accurately, thus giving the observed 
result. This explanation is ad hoc; a better one will have to await more detailed 
study. 

Total Revenue from Gifts. Gift revenue seems to substitute for NIH research 
funding to a small degree in the current period. Over the longer run, however, gifts 
seem to come with NIH funding, suggesting that to some degree NIH funding is 
looked upon as an indication of merit by other donors. With AAMC data, an increase 
in lagged NIH research funding of $1000 brings a drop in gift revenue of about $155. 
Decreases in NIH funding have no discernible effect. 

When we use IMP AC data, this negative relationship continues to hold (now in 
the current period, because of the nature of the data). An increase in NIH research 
funding of $1000 brings a loss of gifts of about $115. Once again a loss in NIH funding 
seems to have no effect. A lagged NIH funding increase of $1000 brings $143 in 
additional gift income. Lagged decreases seem to have no effect. 

Changes in NIH training grants show the curious behavior of seeming to pro- 
mote higher gift revenue no matter in which direction the change is. An increase 
of $1000 brings in $403 in extra gifts and a decrease of similar size brings $315 in 
extra gifts. 

The picture seems to be one of institutional action in the face of changed circum- 
stances. When research grants decrease, gifts are sought, but in the long run gifts 
are attracted by the same qualities that attract NIH grants. Any change in teaching 
funds appears to encourage the seeking of new gift income: attempting to make up 
for shortfalls and being encouraged by increases. Perhaps in unchanging circum- 
stances medical schools become complacent in seeking gifts, but it is more likely that 
this finding is a peculiarity of the data used. 

Conclusions 

It is difficult to summarize the results briefly presented above in any convenient 
way. Changes in NIH research grants seem to exert only a mild influence on reve- 
nues received from other sources, and the effects are usually the result of increases 
in NIH funding. Revenues from sources other than NIH apparently do not make up 
very much for shortfalls in NIH funding and respond only slightly to increases in 
NIH funding. With training grants the picture is somewhat different. NIH funds 
apparently are being substituted for by funds from nonfederal sources. 

In all cases the ratio of changes in funds from alternative sources to changes in 



76 



NIH funds is far less than one to one. There is simply no evidence that changes in 
NIH funding have a major effect on funding from other sources. 

Two caveats should be kept in mind. First, none of the equations estimated fits 
very well. In no case have we been able to explain more than 25 percent of the 
variance in funding from other sources. Given the widely divergent natures of 
medical schools in the United States, it is unlikely that any model could accurately 
characterize such a disparate group of institutions, particularly since we examined 
data on changes from one year to the next. These data are influenced by many 
factors beyond those considered in this simple model. Even with the poor fit of the 
equations, some patterns of behavior stand out. All results presented here, unless 
otherwise noted, are statistically significant at the 5 percent level. Relationships not 
reported may in fact exist, but they cannot be established with the data available. 

The second caveat is that inasmuch as there is a large cross-sectional component 
to this analysis, one must be careful in extending these results to a situation very 
different from that of the last seven years. To be more specific, medical schools often 
compete for revenue from many of the sources considered here. In many cases if one 
school receives money, another does not. A cut in NIH funding to a particular school 
could have very different effects depending on whether similar cuts were also being 
experienced by other schools. The unraveling of this interrelationship among 
schools will have to await a more elaborate analysis. 



RESOURCE ALLOCATION DECISIONS WITHIN ACADEMIC 
MEDICAL CENTERS 

The Model 

The preceding part of this section examines the relationships among the various 
sources that support a medical school, but it would be incorrect to infer from that 
discussion of aggregate funding that all or even most of the school's resources, much 
less the center's, are under the control of the central administration. To be sure, the 
individual with the broadest responsibility for budget, program, and policy decisions 
is the medical school dean or vice-president. The parent university, as well as most 
outsider organizations and agencies, provides support to the center, at least nomi- 
nally, through the dean's office. However, as any medical school dean would hasten 
to point out, having broad policy responsibility and having money pass through his 
hands do not mean that he controls the programs to which the money is allocated. 
Real control of resources is highly decentralized. 

Control over programs stems in part from control over the use of funds that are 
available to a center. This control, in turn, depends on the dean's relationship to the 
provider of money. The continuum of relationships runs from those in which the 
dean's office serves merely as a conduit through which predetermined amounts of 
money are passed to predesignated recipients (e.g., research project grants), to those 
in which the dean acts as the agent of the provider in deciding how money is to be 
used (e.g., general research support grants). 

Control over programs also depends on the dean's obligation to potential recipi- 
ents of the resources he controls. In some instances it is implicitly, if not explicitly, 
the dean's responsibility to provide the resources necessary to meet specific require- 



77 

ments, independently of program action. An example is the base salary for tenured 
faculty. In other cases the dean's responsibility is directly limited to the recipient's 
participation in a program that is the subject of an agreement between the medical 
school and the outside funding agency. An example is a research associate (a term 
employee of the center) whose work is funded by a grant from the Health Resources 
Administration. At the other end of the continuum of obligations is the situation 
where someone in the center requests some money over which the dean has complete 
control for the purpose of starting a new program or, in the case of a department 
chairman, hiring a new faculty member. 

The dean's resource allocation process may be viewed as a system for mixing 
funds from different providers to support activities that each approves to meet the 
dean's obligations. He seeks a funding mix that will promote the programs of the 
center that are consistent with his policy. The mixing process provides a means for 
the dean to meet prior obligations, to use funds for purposes specified by the pro- 
viders, and to exercise some marginal control over programs with his scarce institu- 
tional funds — that is, the funds a provider has not already earmarked for a particu- 
lar center function or individual. (Examples include university general funds and 
federal capitation grants.) 

Our description of the dean's budget suggests that many factors influence his 
decisions on how to allocate the scarce resources that are not already earmarked by 
funding organizations for particular medical center activities. It is not feasible to 
specify a mathematical model that takes account of all these factors since many of 
them are not easily measured. However, by means of simple models it is possible to 
describe some budget outcomes that the dean influences. The most easily interpreta- 
ble of these relate to individual department budgets. The department is probably the 
most useful organizational unit to analyze in terms of the effects of federal biomedi- 
cal research funding because many center activities are organized along departmen- 
tal lines. Examples are the graduate programs in the basic sciences, the intern and 
residency programs, the major services of the teaching hospitals, and much of the 
biomedical research outside the large centers or research institutes. 

A question of great significance to federal research policy is the effect of receiv- 
ing research grants on a department allocation of institutional funds. The answer 
is important from several standpoints: 

• If the dean reduces institutional funds when a department receives more 
money from NIH or offsets reductions in NIH funding with institutional 
funds, the effect of earmarking federal funds for research is mitigated. 

• If the dean increases a department's allotment of institutional funds when 
it gets increased NIH funding, then the federal government may be getting 
more for its money than it spends, and it is almost certain that research 
funds are not "subsidizing" medical education, at least at the margin. 

• If a dean does not compensate for losses in NIH money with institutional 
funds or even cuts back on those allocations, then departments are ex- 
tremely vulnerable to changes in NIH support. 

To analyze the effects of research funding on the dean's allocation of institution- 
al funds to a department, a regression equation is fitted for each department of nine 
medical schools in our sample often. The data used are both cross-section and time 
series, in that the observation is a department budget year. The dependent variable 



78 



is that year's allocation of institutional funds. The explanatory variables include 
last year's allocation of institutional funds (both because we expect that a chairman 
will use that as a point of departure for bargaining with the dean and because we 
expect that real commitments to tenured faculty will impose year to year continuity 
in the institutional budget. The other explanatory variables are year to year changes 
in earmarked funding from outside resources (e.g., NIH research grants) and clinical 
department generated funds from patient care activities. Since we expect that a 
dean would treat a department's gain in outside funding differently from its losses, 
we have decomposed earmarked funding into two variables, one for year to year 
increases and the other for year to year decreases. 

The data for the analysis were collected from the financial offices of the individu- 
al medical schools. We have based our analysis on two kinds of department budget 
data: (1) the total department budget as defined by the medical school business office, 
and (2) the total department budget for faculty compensation as extracted from 
individual faculty compensation records. Where they are available, we chose to use 
the latter kinds of data. The data for total department budgets present problems 
because the financial boundaries of departments vary across departments within an 
individual school, across schools, and over time. 

The number of years of comparable department budget data varied across the 
schools in our sample. In all of the examples for which results are reported, we had 
data for at least four consecutive years through the academic year 1974-1975. 

Our fund source data were in most cases collected in very disaggregated form — 
that is, by account number. We then developed a common accounting framework to 
fit all institutions' data and mapped the new data into that framework. Each result- 
ing category specified the source of funds, the designated use, and the funding 
instrument. For example, NPG signified an NIH (N) program-project (P) grant (G). 
However, because the sample size for a regression is limited by the product of the 
number of departments in a school and the number of years of data less one (for 
differences), we had to aggregate budget data into broad classes, which became the 
variables for our equation. These are shown in Table 32. 

Analysis and Results 

The results of the multiple linear regression analysis for nine academic medical 
centers are presented in Table 33. The dependent variable is the amount of institu- 
tionally controlled funds (university general funds and federal institutional support 
etc.) allocated to a particular department in a particular year. The coefficients for 
each of the explanatory variables are presented in the table. Since the first variable, 
LGNFNDPY, is the department's allocation of institutionally controlled funds in 
the previous year and the remaining variables, RESTRNINC etc., are year to year 
differences in categories of funding, coefficients for these two classes of variables 
require different interpretations. 

The coefficients for LGNFNDPY may be interpreted as the base budget for 
institutionally controlled funds as a proportion of last year's level. That is, a depart- 
ment chairman in School C may be viewed as starting with 93 cents on the dollar 
from his last year's budget of general funds. Changes from this base depend on his 
department's or the school's success in generating other funds and on factors not 
accounted for in our model. 



79 



Table 32 



VARIABLES FOR INSTITUTIONAL BUDGET ANALYSIS 



Variable in Regression 

Institutional Funds (LGNFND, 
LGNFNDPY) 3 



Federal Research and 
Training Funds , 
(RESTRNINC, RESTRNDEC) 

Year-to-year difference in nonfederal 
research and research training funds 

(NFRESTRND) 

Year-to-year difference in other HEW 
earmarked funds (HEWPGMD) 



Year-to-year difference in patient 
care revenue (PATCARD) 



Year-to-year differences in other 
funds (OTHERD) 



Funding Sources Included 

University general funds, tuition 
state appropriations , endowment , 
capitation, etc. 

HEW, NSF, and other federal agency 
funds for biomedical and behavioral 
research and research training. 

Foundations, industry, other 
private grants and gifts. 



Special project grants, Regional 
Medical Programs, and other HEW 
programs that are not directly 
related to biomedical research. 

Revenue from individual and 
clinical faculty group practice, 
contracts with hospitals, VA 
hospitals, and other patient 
care activities. 

Sources not otherwise 
specified. 



LGNFND is the current year allocation and is the dependent variable in all 
regressions. LGNFNDPY is the corresponding amount for the previous year and is 
an independent variable in all regressions. 

Federal research and research training funds are decomposed into two 
variables; RESTRNINC and RESTRNDEC. RESTRNINC is used when there is an increase 
in amounts between two successive years; when there is a year-to-year decrease, 
it takes the value of zero. RESTRNDEC is used when there is a year-to-year 
decrease, and it takes the value of zero for increases. 



The coefficients on the variables (RESTRNINC, RESTRNDEC, NFRESTRND, 
PATCARD, HEWPGMD, and OTHERD) are conceptually more easily interpreted. 
They are the estimated effect of a change in each category of support between two 
years on a particular department's allocation of institutional funds. However, in 
most cases they are not statistically significant at the .05 level. For an example 
where the regression coefficient is statistically significant, see School D. A depart- 
ment in that school may expect to get $.36 more in institutional support for faculty 
salaries for every dollar increase in research and training grant dollars it applies 
to faculty salaries. 

The regression results showing coefficients very close to 1.0 for LGNFNDPY and 
highly significant statistical relationships were expected and reflect the strong year 
to year continuity in institutional budget allocations to departments and in budgets 
more generally. Last year's budget is almost always an excellent predictor of next 
year's budget, no matter what the setting. What is perhaps surprising is that the 
variable LGNFNDPY alone is not sufficient to explain more of the variance in these 
budget allocations than our entire model explains. 

Although there is a significant positive coefficient for RESTRNINC in only one 
school, the fact that the signs are positive indicates that increases in NIH and other 



80 



CI 


^ 






r. 


W 


CU 




■\i 


— 


*i 




u 




u 




n> 


u 


Bl 


m 


4-1 


0) 


fc 


T) 



Q C U h 



<U 0) O C 
U TJ U -H 

<u 1) tq c 



o o 



\D CTi CT* tTl 



O O 

o o 



o o 








00 vO 


(N m 


vD \£> 


en 


<T 


a\ o 


c^ o 


o o 


cr> O 


o 


O 



<u 


m 


.* 


— < 


(fl 


1 


R 


-st 




r— 


(fl 


en 


-j 


— / 



&0 aJ -H u 



■P iH 



ai 4-i <4h 



81 



research funds are matched by these medical centers' own funds, that it may cost 
the institution something to participate in research programs. (A sign test shows the 
effect to be positive and significant at the .05 level.) These results tend to contradict 
claims that institutions use research funds to "free up" their own resources for other 
purposes. Only in School H is there evidence (a minus sign) of such budget behavior 
and this is not statistically significant. 

The absence of significant negative coefficients for RESTRNDEC indicates that 
institutional funds are not used consistently to replace a department's loss of re- 
search funds. Indeed, School A appears to reduce the allocation of institutional funds 
substantially when a department loses research funds. This suggests that depart- 
ments are left to fend for themselves in the face of cutbacks in research support and 
probably reflects a general shortage of institutional resources. 

There is no consistent pattern among the coefficients for other categories of 
funds. The significant positive coefficients for patient care revenue in Schools E and 
F probably reflect recycling of department-generated practice revenue through a 
"dean's fund" (institutional funds) back to the department. 

Conclusions 

The results of our analysis indicate that academic medical center departments 
function as entrepreneurial units whose fortunes depend in substantial part on their 
ability to generate funds from outside sources. NIH and other public and private 
research funding agencies are important sources of department-generated funds, 
and practice earnings are of growing importance to all clinical departments, particu- 
larly those of high-earning specialties. 

The central administration of a medical center appears to exercise only limited 
control over total department budgets, at least in the short term. There may be some 
asymmetry in the central administration's budget behavior with respect to increases 
and decreases in department research and research training funds, but the asymme- 
try is different from what one might expect. That is, it appears that in most institu- 
tions, a department may obtain more institutional funds by increasing its research 
support. However, a department's loss of research funds does not appear to have a 
significant effect on its allocation of institutional funds. This indicates that individu- 
al departments may be quite vulnerable to research funding cutbacks. 

Another inference that may be drawn from our analysis relates to the question 
of whether research funds "subsidize" the education programs of medical centers. 
There is no evidence of such subsidies in our analysis. It appears that in most of the 
schools in our sample a department's research success may enable it to attract 
matching institutional funds. We find no evidence that research funds supplant 
institutional funds that are generated by, or would normally be used exclusively for, 
the training of undergraduate M.D.s. That is not to say that the character of M.D. 
education programs is not influenced by the presence of a research effort. 



V. THE INFLUENCE OF FEDERAL BIOMEDICAL 
RESEARCH FUNDING ON RESEARCH AT UNIVERSITIES 



This section examines the characteristics of the biomedical and behavioral re- 
search funded by NIH and ADAMHA and performed in an academic setting: gradu- 
ate departments of universities, colleges of arts and sciences, schools of medicine and 
affiliated hospitals, and other health professions schools. The analyses described in 
the previous sections have sought to differentiate the effects of the research program 
from the effects of other factors that combine to shape the institutional character- 
istics (e.g., education programs, budget decisions) of the academic medical centers. 
This section focuses only on the research programs, ignoring the possible effects of 
education and health career programs on the research activities of the universities. 
We examine the relationships among federal research programs (e.g., ADAMHA, 
NIH, NIMH, NCI), the various segments of the university community that receive 
funds from each program, and the scientific content of the research that has been 
funded. 

Our first step is to describe the research that is being performed under federal 
sponsorship. From the point of view of the federal government, an important attri- 
bute of biomedical and behavioral research is the agency, Institute, or program 
within the Institute that sponsors the research. Identifying the sponsor enables us 
to describe, at least roughly, the disease entity or normal process that will be better 
understood as a result of the research. Budget allocations among programs are made 
on the basis of the health problem being studied. The review processes for NIH and 
ADAMHA are administered separately. Differences among the federal programs 
may have different effects on the different parts of the university. 

From the scientist's point of view, scientific activity is described less by the 
sponsoring agency than by the content and methods of the research. Although some 
sponsoring agencies do consistently fund certain kinds of research, the agency alone 
is not an adequate description of the research. There is not enough detail to describe 
the variety of biomedical research. In some areas of science, research that is very 
similar in its methods, knowledge base, and scientific goals is sponsored through 
programs of different institutes. Changes over time in the sponsoring agency do not 
always signal real changes in scientific activity because there have been changes in 
the perception of the areas of basic research that are relevant to certain diseases. 
In addition, it might be possible for an investigator to influence the assignment of 
his application to an Institute by adding a disease-oriented window dressing to his 
application, without changing the scientific content of his work. To avoid these 
problems, scientific activity has been classified based only on its scientific content, 
and not in any way on the structure of federal programs. This classification permits 
exploration of thp effects federal program decisions have had on scientific activity. 

The data consist of all applications in FY 1971 to 1975 for traditional single 
investigator research project grants 1 from educational institutions: medical schools, 
affiliated hospitals, other health professions schools, graduate departments, and 
colleges of arts and sciences. 

1 Coded R01 on the IMPAC file. 

82 



83 



Federal program characteristics are used to present the effect of current review 
practices and funding levels for NIH and ADAMHA on support of the different 
components of the university. Then the typology of the scientific activity performed 
under NIH research project grants is used to describe how federal funding priorities 
affect the scientific characteristics of research performed in institutions of higher 
education. 



FEDERAL PROGRAM CHARACTERISTICS 
Assignment of Applications to Institutes 

Changes in the relative funding levels of the various federal programs will have 
a greater effect on the parts of the university that are most heavily involved in the 
programs being enlarged or cut back. There are significant differences in the aca- 
demic settings of research among the institutes of NIH and ADAMHA (see Table 
34). 

Applications assigned to NIGMS are much more frequently from a university 
science department than from a medical school basic science department. Appli- 
cations for NHLI and NIAMDD are most frequently from clinical departments of 
medical schools, and applications from basic science departments are frequently 
from the medical school side of the university. The situation for basic science at 
NICHD is reversed, probably because of population and psychology studies. How- 
ever, applications come to NCI from each component of the university in the same 
proportion as total NIH applications. 

ADAMHA receives 75 percent of psychiatry department applications but only 
10 percent of applications from the rest of the university and medical school depart- 
ments. Within ADAMHA, departments of the medical school other than psychiatry 
have a higher than expected proportion of their applications going to NIAA and 
NIDA, while university departments are more likely to apply to NIMH. Appli- 
cations from departments of psychiatry and health professions schools go to each of 
the three Institutes of ADAMHA in the same proportion as the total number of 
applications. 

Rate of Disapproval of Applications 

In general, ADAMHA disapproved a larger proportion of applications than NIH. 
NIH disapproved 33 percent of new applications and 1 1 percent of renewal appli- 
cations, while ADAMHA disapproved 52 percent of new applications, 22 percent of 
renewal applications (see Tables 35 and 36). Aside from graduate schools, we find 
that in both NIH and ADAMHA, the rate of disapproval of applications follows the 
same pattern by source, with basic science departments having fewest of their 
applications disapproved. However, in all these cases, ADAMHA applications are 
disapproved much more frequently than NIH applications. The difference in the rate 
of disapproval of applications from graduate schools relative to all applications for 
each agency may be because applications to NIH are likely to be from basic science 
departments, which have a higher approval rate, and applications to ADAMHA are 
likely to be from behavioral sciences departments, which have a lower approval rate. 

The rate of disapproval of applications varies among the study sections in both 



84 



Table 34 



PERCENTAGE OF COMPETING RESEARCH PROJECT GRANT APPLICATIONS 
FROM COMPONENTS OF UNIVERSITIES BY INSTITUTES 
OF NIH AND ADAMHA 



Graduate 
Departments 
Basic Science Clinical Science Psychiatry Health and Schools 
Departments of Departments of Departments of Professions of Arts 
Medical Schools Medical Schools Medical Schools Schools and Sciences 



Institute 




of NIH 




NAI 


28.4 


NIAID 


34.1 


NIAMDD 


29.8 


NCI 


33.0 


NIDR 


13.8 


NIESH 


30.2 


NEI 


13.2 


NIGMS 


26.1 


NICHD 


19.6 


NHL I 


32.1 


NLM 


13.9 


NINCDS 


34.6 


DRR 


14.8 


Total NIH 


28.6 



Number of NIH 
Applications 



8747 



18.8 
30.0 
51.3 
37.5 
7.6 
23.1 
54.1 
13.0 
32.2 
52.2 
26.8 
33.6 
51.9 
35.6 

10878 



2.0 


9.6 


41.1 





6.9 


29.1 


0.1 


4.0 


14.8 


0.3 


5.9 


23.2 





70.0 


8.7 


0.4 


14.5 


31.9 


0.6 


4.2 


28.0 


0.3 


4.2 


56.4 


3.3 


8.3 


36.7 


0.5 


5.1 


10.1 


5.2 


12.1 


42.0 


3.7 


4.4 


23.8 


" 


11.1 


22.2 


1.1 


7.4 


27.4 



324 



2274 



8375 



Institute 


of 




ADAMHA 






NIAA 




17.6 


NIDA 




25.6 


NIMH 




6.1 


Total ADAMHA 


9.5 


Number of 


ADAMHA 




Applications 


424 



24.8 

20.3 

7.0 

10.1 

450 



24.5 


6.0 


27.1 


27.9 


6.5 


19.7 


20.8 


6.4 


59.7 


21.9 


6.4 


52.1 



975 



283 



2313 



85 



Table 35 



DISAPPROVAL RATE FOR RESEARCH PROJECT GRANT APPLICATIONS 
TO NIH BY COMPONENTS OF THE UNIVERSITY 3 



New Applications 



Renewal Applications 



Number of Percentage Number of Percentage 
Applications Disapproved Applications Disapproved 



Basic Science 
Department of 
Medical School 



5664 



28.1 



2894 



9.8 



Clinical Science 
Department of 
Medical School 



7636 



37.2 



3008 



14.2 



Psychiatry 
Department of 
Medical School 



258 



52.3 



59 



10.2 



Health 










Professions 










Schools 


1739 


41.9 


502 


15.1 


Graduate Depart- 










ments and Schools 










of Arts and 










Sciences 


5882 


30.2 


2319 


7.1 


TOTAL 


21179 


33.4 


8782 


10.9 



i) Competing R01 applications from IMPAC file fiscal years 1971-1975. When 
amended applications appear on the IMPAC tape, only the last amendment is 
counted. 

Table 36 

DISAPPROVAL RATE FOR RESEARCH PROJECT GRANT APPLICATIONS 
TO ADAMHA BY COMPONENTS OF THE UNIVERSITY 3 



New Applications 



Renewal Applications 



Number of Percentage Number of Percentage 
Applications Disapproved Applications Disapproved 



Basic Science 
Department of 
Medical School 



299 



41.1 



120 



17.5 



Clinical Science 
Department cf 
Medical School 



331 



46.2 



100 



27.0 



Psychiatry 
Department of 
Medical School 



734 



52.6 



206 



23.8 



Health 

Professions 

Schools 



227 



52.4 



42 



23.8 



Graduate 
Departments and 
Schools of Arts 
and Sciences 

TOTAL 



1772 
3363 



54.2 
51.8 



474 
942 



21.9 
22.4 



Competing R01 applications from IMPAC file, fiscal years 1971-1975. 
When amended applications appear on the IMPAC file, only the last 
amendment is counted. 



86 



NIH and ADAMHA. However, no meaningful grouping of the study sections has 
been found to explain the pattern in disapproval rates. 

Over the time period of FY 1971 to FY 1975, NIH study sections approved more 
applications, both new and renewal, in the later years than in the earlier ones; in 
FY 1971 and FY 1972, an average of 37 percent of new applications and 14 percent 
of renewal applications were disapproved; in FY 1974 and FY 1975, an average of 
29 percent of new applications and 9 percent of renewal applications were disap- 
proved. No trend is observable in the rate of disapproval of ADAMHA applications. 

Rate of Funding of Applications 

The rate of funding of approved NIH applications is the same for applications 
from all of the university components; about 59 percent of approved applications are 
funded. This is not true in ADAMHA. Once approved, ADAMHA applications from 
all parts of the medical and other health professions schools are funded about 80 
percent of the time, while approved applications from the university departments 
and schools of arts and sciences are funded only 66 percent of the time. The disposi- 
tion of applications is summarized in Table 37. 



Table 37 



DISPOSITION OF RESEARCH PROJECT GRANT APPLICATIONS 



Basic Science Departments 
of Medical Schools 

Clinical Science Depart- 
ments of Medical 
Schools 

Psychiatry Departments 
of Medical Schools 



Percent of Percent of Approved Percent of All 
Applications That Applications Applications 
Were Disapproved That Were Funded That Were Funded 



NIH ADAMHA TOTAL NIH ADAMHA TOTAL NIH ADAMHA TOTAL 



22 



31 



45 



34 



42 



46 



23 



31 



46 



60 



59 



52 



HI 



til 



79 



61 



59 



11 



47 



41 



29 



53 



48 



42 



47 



41 



39 



Health Professions 




















Schools 


36 


48 


37 


59 


82 


61 


38 


43 


39 


Graduate Departments and 




















Schools of Arts and 




















Sciences 


24 


47 


29 


59 


66 


60 


45 


35 


43 


Total 


27 


45 


29 


59 


73 


60 


43 


40 


43 



Competing R01 applications from IMPAC file, FY 1971-1975. When amended applications 
appear on the IMPAC Tape, only the last amendment is counted. 



87 

SCIENTIFIC CHARACTERISTICS 

Data Base 

We have chosen to use the scientific classification codes assigned to approved 
applications to NIH for traditional research project grants (ROD from the IMPAC 
file to develop a typology of biomedical science that would describe the subfields 
within this broad field. Unfortunately, no similar classification system is available 
for applications to ADAMHA. 

The scientific classification system describes the research to be performed along 
four axes. The first axis gives up to three codes for discipline and field such as 
biochemistry, physiology, etc. The complete list of codes is found in Table 38. The 
second axis gives up to two codes for the body system or systems under consideration; 
allowable codes are enumerated in Table 39. The third axis describes the research 
materials used for the research. Each code in the third axis describes both the source 
of the research material (e.g., humans, animal), the stage of development being 
studied or used as a source (e.g., infancy, childhood), and whether the research has 
developmental aspects (i.e., the study concerns changes over time or events at one 
stage that cause changes at a later stage). A research material code consists of a code 
from each section of Table 40 (a special code is used when stage of development is 
not applicable — e.g., with research materials or computer) with a binary indicator 
for the existence of developmental aspects. For simplicity in the analysis, we have 
separated the components of each research material code into its three parts and 
used them as if they were separate codes. This should not be a major problem, but 
it does mean that we will treat a grant that studied the infancy period of animals 
and the childhood period of humans the same as one that studied the infancy period 
of humans and the childhood period of animals. The fourth axis is a binary code that 
tells whether the project is drug related. 

It may be worth discussing several available indicators of the kind of research 
that we have chosen not to use. The disease category to which the work is applicable 
is of course a policy-relevant indicator of the kind of research, and the Institute 
assignment of the application gives information about that category. However, an 
investigator may succeed in influencing the Institute to which an application is 
assigned by orienting a basic research proposal toward a specific disease when his 
proposal is applicable to the charter of more than one Institute. Insofar as this is 
true, the Institute assignment of the application reflects not the scientific content 
of the work but rather only the content of the proposal. Another problem with using 
Institute assignment has to do with changes over time in the perception of the 
disease category to which certain fields of basic research are relevant. For example, 
the Cancer Institute supported 37.4 percent of competitive renewals in immunology 
in 1974, whereas it had supported only 11.6 percent of these grants in 1971. In 1974 
NCI took over 68 immunology grants that had previously been funded by NIAID. 2 

Another indicator of the kind of research is the Initial Review Group for the 
application. Here the case is not so clear for either inclusion or exclusion, since the 
study section assignment should reflect the broad category of the research. However, 
the charters of some study sections overlap because of the large numbers of appli- 

2 Herman N. Eisen et al., Final Report of the Immunology and Microbiology Interdisciplinary Cluster, 
October 8, 1975. 



88 



Table 38 

DISCIPLINE AND FIELD CODES ASSIGNED TO APPROVED COMPETING APPLICATIONS 
FOR RESEARCH PROJECT GRANTS (AXIS T) 



SCIENTIFIC DISCIPLINE 



1971 1972 1973 1974 1975 



Total 



1100 


Physics 


9 


7 


13 


17 


10 


56 


1200 


Chemistry 


366 


409 


314 


453 


459 


2001 


1240 


Structural Chemistry of 
















Biopolymers 


351 


286 


188 


223 


247 


1295 


1300 


Biochemistry 


1436 


1741 


1655 


2224 


2376 


9432 


1500 


Pharmacology 


393 


492 


513 


641 


697 


2736 


1600 


Toxicology 


75 


79 


77 


91 


123 


445 


1700 


Physiology 


912 


1064 


1105 


1455 


1529 


6065 


1900 


Nutrition 


97 


100 


94 


124 


142 


557 


2000 


Microbiology, Excluding 
















Virology 


108 


159 


175 


218 


271 


931 


2100 


Parasitology 


42 


64 


67 


59 


72 


304 


2210 


Immuno gene tics 


27 


48 


49 


61 


74 


259 


2215 


Immunochemistry 


145 


189 


189 


237 


254 


1014 


2220 


Immuno pat ho logy 


128 


160 


220 


257 


295 


1060 


2225 


Hypersensitivity 


38 


45 


38 


64 


55 


240 


2230 


Immunotherapy 


6 


33 


47 


63 


91 


240 


2299 


Immunology, Other 


143 


177 


178 


224 


296 


1018 


2300 


Genetics 


374 


424 


452 


530 


581 


2361 


2400 


Cell Biology 


331 


380 


366 


441 


510 


2028 


2500 


Virology 


123 


153 


154 


186 


230 


846 


2600 


Anatomy 


169 


217 


194 


265 


329 


1174 


2700 


Pathology 


299 


353 


437 


503 


552 


2144 


2900 


Biology, Not Elsehwere 
















Classified 


88 


66 


84 


112 


96 


446 


3100 


Social Sciences 


70 


88 


79 


126 


111 


474 


3200 


Psychology 


184 


199 


193 


273 


219 


1068 


3400 


Reproduction, Growth & Dev. 


195 


300 


287 


405 


423 


1610 


3500 


Epidemiology 


16 


42 


33 


64 


57 


212 


3600 


Mathematics 


73 


73 


83 


95 


79 


403 


3700 


Information and Communication 
















Sciences 


39 


39 


34 


39 


32 


183 


3800 


Bioengineering and Instrumen- 
















tation 


117 


166 


177 


209 


194 


863 


3900 


Biomaterials 


14 


20 


14 


12 


28 


88 


4100 


Environmental Health Sciences 


17 


19 


16 


20 


29 


101 


4200 


Health Sciences and Services 


11 


11 


14 


15 


20 


71 


4310 


Biological Resources 


6 


3 


2 


5 


5 


21 


.4320 


Animal Production & Facilities 


6 


3 


4 


4 


4 


21 


4400 


Clinical Medicine, General 


32 


96 


93 


81 


96 


398 


4505 


Anesthesiology 


5 


10 


12 


12 


15 


54 


4510 


Oncology 


53 


57 


93 


122 


161 


486 


4514 


Radiology 


15 


42 


62 


89 


85 


293 


4515 


Transplantation, Other Than 
















Transplantation Immunology 


64 


68 


53 


75 


75 


335 


4520 


Surgery 


63 


138 


144 


174 


139 


658 


4521 


Trauma 


20 


34 


41 


30 


41 


166 


4525 


Denistry 


19 


21 


16 


10 


45 


111 


4530 


Hematology 


83 


88 


70 


89 


92 


422 


4550 


Opthalmology 


37 


24 


17 


19 


21 


118 



89 



Table 39 

BODY SYSTEM CODES ASSIGNED TO APPROVED COMPETING APPLICATIONS FOR 
RESEARCH PROJECT GRANTS (AXIS II) 



1971 1972 1973 1974 1975 



Total 



100 Whole Body 

110 Oral and Dental 

120 Body Cavities and Fluids 

130 Lymphatic, Hematopoietic and 

Reticule endothelial Systems 

140 Skin and Membrane 

150 Connective Tissues 

160 Muscle 

170 Nervous System 

190 Sensory Systems, Other than 

Visual Systems 

200 Eye and Visual Systems 

220 Endocrine & Exocrine Systems 

240 Circulatory System 

260 Respiratory System 

270 Gastrointestinal System 

290 Urinary System 

300 Reproductive System 



355 


346 


340 


411 


457 


1909 


51 


94 


62 


80 


113 


400 


551 


632 


627 


831 


922 


3563 


181 


200 


217 


271 


327 


1196 


79 


163 


159 


225 


212 


838 


98 


147 


120 


142 


159 


666 


131 


124 


124 


198 


217 


794 


362 


425 


423 


619 


659 


2488 


96 


100 


77 


150 


118 


541 


127 


160 


182 


240 


202 


911 


160 


237 


270 


340 


400 


1407 


338 


349 


424 


449 


494 


2054 


103 


134 


130 


189 


176 


732 


308 


439 


354 


472 


496 


2069 


192 


227 


213 


293 


258 


1183 


160 


189 


188 


268 


319 


1124 



Table 40 

RESEARCH MATERIAL CODES ASSIGNED TO COMPETING APPLICATIONS 
FOR RESEARCH PROJECT GRANTS (AXIS III) 





S OF DEVELOPMENT 


1971 


1972 


1973 


1974 


1975 


Total 


STAGE 














11 


Preconception 


13 


14 


15 


16 


16 


74 


12 


Prenatal 


112 


125 


112 


143 


131 


623 


13 


Perinatal 


59 


70 


74 


76 


81 


360 


14 


Infancy (First Year of Life) 


44 


47 


43 


72 


56 


262 


15 


Childhood 


91 


72 


73 


88 


70 


394 


16 


Adolescence 


17 


20 


18 


22 


20 


97 


17 


Adult 


244 


265 


262 


367 


352 


1490 


18 


Aged 


13 


22 


11 


31 


38 


115 


19 


Combination of 3 or More 


152 


253 


268 


326 


345 


1344 


20 


Life Span 


57 


39 


43 


41 


59 


239 


21 


Pregnancy 


9 


40 


22 


28 


41 


140 



RESEARCH MATERIALS 



1 Individuals, Human 

2 Individuals, Human & Animal 

3 Individuals, Animals 

56 Microorganisms, Including 
Virus and Protozoa 

60 Plant 

85 Computers 

90 Other 

98* Biological Subsystems: 

Tissues, Organs, Tumors, 
Body Fluids, Cells and 
Cell Lines 

TOTAL NUMBER OF APPLICATIONS 



387 516 600 668 677 

255 313 222 334 362 

1555 1809 1795 2376 2506 

544 693 636 763 847 

65 54 53 64 72 

94 105 121 161 133 

341 451 378 499 523 



1522 1727 1912 2582 2889 
3273 3882 3711 5103 5358 



2848 

I486 

10041 

3483 
308 
614 

2192 



10632 
21327 



Temporary category summarizes all codes which cannot be uniquely mapped 
between systems used in different time periods. 



90 



cations in a field. Study sections have been added and deleted over time. In addition, 
the content of the work reviewed by a study section might gradually change over 
time without a change in the title of a study section. A minor problem is that study 
section members' applications are not assigned to their own study sections, even 
though in many cases that would be the most scientifically appropriate selection. We 
produced two sets of clusters, one including study section data and one without it. 
Unfortunately, when the study section was included, this variable appeared to 
dominate the clustering assignment, leading to fears that changes in the procedures 
used for study section assignment might overwhelm our analysis of changes in 
scientific content. Consequently, we have used the set of clusters without study 
section data in all our analyses. 

Tables 38-40 list the number of competing R01 applications from selected compo- 
nents of institutions of higher education that received each code in fiscal years 1971 
through 1975. When amended applications appeared on the IMPAC tape, only the 
last amendment was included in these tables and all subsequent analysis. In general, 
codes are not assigned for disapproved applications, so the population consists of 
only approved applications — i.e., those with enough scientific merit to be funded if 
the money is available. Unfortunately, in FY 1971 codes were not assigned to 189 
approved applications, and in FY 1975 they were not assigned to 314 applications. 
However, this should not be much of a problem as it is less than 6 percent of the 
applications in each of these two years, and the grants without codes appear to be 
spread randomly over Institute and IRG. 

Some trends are evident in these three tables and in Table 41, which shows the 
study section assignments for the same set of applications. One example is in the 
increasing assignment of the Immunology codes (2210-2299); they were assigned 
only 487 times in 1971, but 1065 times in 1975. This 119 percent increase is quite 
large compared with the 64 percent increase in number of approved applications. 
However, as we shall show, looking at research one attribute at a time does not 
provide enough information to enable us to pinpoint the changes that are taking 
place. 

Methodology 

We have attempted to relate each grant to an identifiable subfield of biomedical 
research. Using one code or one axis is not sufficient for this purpose; it is necessary 
to use the entire combination of codes assigned to each grant. For example, the 
biochemistry code appears on about 44 percent of applications, but one would clearly 
want to distinguish two grants that received the biochemistry code as belonging to 
separate subfields if one grant received the codes for pharmacology and nervous 
system along with biochemistry, and the other received the codes for physiology and 
the reproductive system. 

Using the entire set of codes assigned to each application presents the opposite 
problem from the use of just a single code. Since each grant may receive as many 
as three codes for discipline and field, two codes for body system, two 5-digit codes 
for research materials, and a code for whether it was drug related, there is such a 
detailed description of each project that most are unique and therefore not amenable 
to analysis. Out of our file of 21,000 competing applications, there are over 11,000 
unique descriptions when all the codes are considered. Nevertheless, some sequences 



91 



Table 41 

INITIAL REVIEW GROUPS FOR COMPETING APPLICATIONS 
FOR RESEARCH' PROJECT GRANTS 



1971 1972 1973 1974 1975 



Total 



AFY Applied Physiology and 

Bio engineering 

ALY Allergy and Immunology 

BBCA Biophysics & Biophysical 

Chemistry A 

BBCB Biophysics & Biophysical 

Chemistry B 

BCM Biomedical Communications 

BIO Biochemistry 

BM Bacteriology & Mycology 

CBY Cell Biology 

CMS Communicative Sciences 

COM Computer & Biomathematical 

Sciences 

CVA Cardiovascular & Pulmonary 

CVB Cardiovascular & Renal 

DBR Developmental Behavioral 

Sciences 

DEN Oral Biology & Medicine 

EDC Epidemiology & Disease Control 

END Endocrinology 

ET Experimental Therapeutics 

EXP Experimental Psychology 

GEN Genetics 

GMA General Medicine A 

GMB General Medicine B 

HED Human Embryology & Development 

HEM Hematology 

1MB Immunobiology 

MBC Microbial Chemistry 

MBY Molecular Biology 

MCHA Medicinal Chemistry A 

MCHB Medicinal Chemistry B 

MET Metabolism 

NEUA Neurology A 

NEUB Neurology B 

NTN Nutrition 

PC Physiological Chemistry 

PHRA Pharmacology 

PHY Physiology 

POP Population Research 

PTHA Pathology A 

PTHB Pathology B 

RAD Radiation 

REB Reproductive Biology 

SGYA Surgery A 

SGYB Surgery B 

TEC Clinical Trials Review (NHLI) 

IMP Tropical Medicine and 

Parasitology 

TOX Toxicology 

VIS Visual Sciences 

VR Virology 



7 


11 


13 


30 


42 


103 


87 


109 


111 


128 


142 


577 


111 


119 


112 


130 


166 


638 


104 


107 


105 


138 


166 


620 


8 


4 


9 


6 


7 


34 


154 


158 


145 


199 


228 


884 


65 


79 


101 


109 


157 


511 


81 


113 


95 


123 


92 


504 


64 


81 


58 


115 


89 


407 


29 


29 


40 


42 


31 


171 


60 


76 


80 


115 


119 


450 


63 


65 


70 


98 


129 


425 


29 


32 


30 


29 


33 


153 


46 


73 


55 


59 


92 


325 


22 


32 


25 


42 


44 


165 


93 


119 


125 


153 


170 


660 


47 


76 


90 


91 


103 


407 


67 


76 


54 


88 


75 


360 


106 


145 


162 


167 


175 


755 


47 


76 


64 


83 


94 


364 


65 


90 


77 


110 


82 


424 


50 


61 


72 


71 


72 


326 


94 


109 


97 


102 


107 


509 


88 


95 


102 


149 


178 


612 


123 


117 


113 


161 


176 


690 


80 


100 


83 


123 


128 


514 


55 


93 


86 


105 


97 


436 


79 


93 


71 


129 


90 


462 


115 


125 


106 


132 


146 


624 


71 


64 


72 


97 


107 


411 


65 


89 


80 


156 


141 


531 


33 


34 


26 


29 


63 


185 


146 


157 


146 


199 


199 


847 


69 


66 


78 


98 


99 


410 


112 


102 


97 


145 


153 


609 


20 


43 


57 


86 


82 


288 


85 


70 


72 


132 


86 


445 


68 


71 


87 


96 


91 


413 


44 


55 


79 


95 


107 


380 


53 


75 


76 


108 


138 


450 


40 


67 


55 


64 


62 


288 


63 


72 


68 


74 


59 


336 





6 


22 


21 


5 


54 


39 


60 


60 


51 


h5 


277 


38 


31 


i5 


50 


70 


224 


75 


107 


126 


1 70 


162 


640 


99 


1 17 


12 1 


145 


1 76 


680 



92 



of codes appear repeatedly on different applications; and, although these sequences 
may not completely describe any single project because additional codes were at- 
tached to each application, they describe the research subfield to which the appli- 
cation can be assigned. For example, the sequence of codes for anatomy, physiology, 
the nervous system, and animals as a research material source occurs frequently 
and describes a subfield within the neurosciences. 

We start with the hypothesis that projects having many of the same codes are 
more similar than projects having only one or no codes in common. We wish to group 
our set of applications into clusters such that each application is more like the set 
of grants in its own cluster than it is like the set of grants in any other cluster. To 
do this we define a function that describes the distance between any two applications 
as the total number of codes that describe either application but do not describe the 
other application. Thus two grants with identical descriptions have distance zero 
between them, and two grants that have all the same codes except for one each have 
distance two between them. We then assign each grant to the cluster that has the 
minimum distance between itself and the other grants in the cluster. 

Most conventional clustering programs require a matrix of the distances be- 
tween each pair of objects to be clustered. Such algorithms are well suited to data 
bases of 50 to 200 data points; but they become quite cumbersome, if not impossible, 
to handle when the data points are as large as even a thousand. Since we must 
cluster at least several thousand applications in order to get a rich description of 
scientific content, we had to develop a new clustering algorithm. To make the 
problem manageable, we calculate and store only distances from object to cluster 
center, not distances from object to object. 

Clusters of Applications by Scientific Field 

Table 42 describes 50 clusters that were produced by our algorithm. Each cluster 
represents a research area that contains at least several hundred research projects. 
To name these clusters, we took advantage of a detailed IMPAC coding system that 
was used on our IMPAC file in 1971 through 1973. For example, most projects with 
the body fluid code use the blood system, and most of the circulatory system codes 
refer to the cardiovascular system. In Table 42 we list only the codes from the 
scientific classification system that appear on at least one-third of the grants in the 
cluster. Most of the clusters can be neatly classified as a subset of one of the research 
areas of the interdisciplinary cluster panels convened by the President's Panel for 
Biomedical Research. Clusters 1-3 are subsets of the neurosciences; clusters 4 
through 10 are from microbiology and immunology; clusters 11-13 are from devel- 
opmental biology; clusters 14-22 are from the tissue and organ system; clusters 23-27 
are from the pharmacology, substance abuse, and toxicology cluster; cluster 28 is 
within social and behavioral development; clusters 29-31 are in the behavioral 
sciences; clusters 32-47 are within biochemistry, molecular genetics, and cell biolo- 
gy; cluster 48 is within the epidemiology, biostatistics, and bioengineering cluster. 
The last two clusters are miscellaneous and contain grants that don't fit neatly 
anywhere else. The two interdisciplinary panels that do not emerge with separate 
clusters are nutrition and communicative sciences. There are few applications in 
these areas relative to many of the other panels. Their applications are scattered 
among several of our clusters. 

When we used the study section assignment as an additional input to the cluster- 



93 



fl> 


[fl 


i 


() 


fx 


(0 D. 


u 




CD o 


(1) 


rt 


00 h 


- 


-H 


« 01 

4-J > 




c 


tn cu 



CD Q. O 
E < u 



w ^ Q 

< Y* H 

HUH 

z z z 



^~. 


,-^ 


O 


^-, 


^-s 




en 


O 


ci 


rH 


o-> 


^o 


H 


CC 


U"l 


o 


o 


O 


m 


(T. 


o 


o 


O 


-H 


CT» 


vn 


r~- 


CN 


rg 




CN 


H 


CM 


CN 


CM 



O *-% /-i 
O CO o 



. *-> o --- ~ 

CN o cm in 

•n H ffi CO 



o o o o o o 

O CM O OC^ 

(N OJ m CM O CM 

CM CN CM CM CM CM 



~ 


s-^ 


o 


m 


H 


-j. 


~ 


(T\ 


O 


O 


CN 


CN 



o o 



co E 


O E 


ME--. 


?^ ai 


•H CU 


O d) >> 


j= *j 


co 4-1 


H 4-1 E 


x, to 


>i en 


O CO o 


>. 


.n ?s 


■H >! 4-1 


T3 CA1 


ex en 


H CO flj 



94 



1 3j 



c -a 

•H tH 

iH QJ 

■H fe 

u 

» -o 

■H C 



o o 

O CM 



o -— 
O r- 

i-l Q\ 



i CO 



ON 3C 



\£) in 

o o o c 

o o c c? 

r~- m oj in 



O O 
O O 



o o 
o o 



o o o 

(MOO 
in r» co 

<■ h n 



o o 
m o 
m m 



If) 


i-l 


O 


^ 


[fl 




■C 


P-. T3 


Pu 


W 


c 
CI 


>, 


0) 




l- 


r 


iH 


u 


■H 


cfl 


{/) 


M 






U 


*H 


R 


O 


C 


0) 


T3 


-H 



(VI 


4J 


O 


(0 


— 1 


>, 


n 


-' 


■H 




U) 


u 


>> 




G 


<-1 


(l. 


3 


-n 


to 


(1) 


m 


u 




cfl 


o 



o 


co E 


■H 


o -■-< 


w 




>> w 


co < 


J= QJ 


o 


0- T\ 


.-H W 





ai 


ca a) 






«-H -H 


<fl 


"O 


oj -a 


u 


3 


OS 3 








c 


C/3 


oc en 



95 



0) 


QJ QJ 


■H TD 


CO 4-1 


4-1 C 


oj cd 


TO U 


at s 



< CO 

03 CO 



9 £ 



■-( 0) 


o f> 


cd ^o in 






CT. CO ^T> 








u 


o o 


o o o 




o o 


o o o 


H C 


m m 


^o m ro 



o o 
o o 

en ro 



o o 
<r o 

OJ CO 



u cc 

01 
(X 00 



96 





u ai 




CQ <-J 


co m 


CQ --, 



Q t/) 

i-i s 



O O O 
-J- O O 
n oj n 



O O 
o o 

ro en 



o o 
o o 



•s-r 


o 


o 


O 


o o 






o 


o 


o o 


in 


m 


cr> 


<a 


co r- 


sr 


-J 


CM 


CO 


n ih 



01 E 
E 01 

i-t C 



97 



r-i V 


r- f-t 


a, -h 


ov in 


H tn 




O 


o o 


m -a 


O <-i 


H C 


r- m 


o ra 


cni -^3- 



98 



ing algorithm, several of the clusters were similar to those shown here. However, 
many of the clusters came to be denned almost exclusively by the study section to 
which applications were assigned and were much less homogeneous with respect to 
the codes of the scientific classification system. Several times grants that were 
identical in all attributes but study section were assigned to different clusters. One 
of the advantages of the clusters with study section assignment was that two sepa- 
rate clusters emerged for the communication sciences area, one for visual research, 
another for sensory perception. In our current system, these grants are linked with 
other grants based on their discipline and research materials (i.e., biochemistry or 
physiology or psychology; clinical studies or animal studies), rather than on the body 
system (eye or other sensory system) as in the alternative set of clusters. 

Our clusters show very clearly the interdisciplinary nature of most biomedical 
research. Much research is taking place at the boundaries of two of the traditional 
fields of bioscience: 31 of these 50 clusters require at least two discipline and field 
codes to describe them. For example, a large part of the surgical research over this 
time period involved the circulatory system (mostly cardiovascular surgery). 

The Initial Review Groups to which most of the applications in each cluster were 
assigned are also shown in Table 42, although they were not used as attributes to 
cluster the grants. Although there is a one to one relationship between a few clusters 
and IRGs, applications in other clusters were split among several IRGs. 

The Institute assignment of the largest fraction of applications in each cluster 
also is shown in Table 42. Only a few of the clusters are clearly related to only one 
Institute, with many of the scientific fields being supported by several Institutes. 
Applications to the Eye Institute and the Dental Institute are small portions of 
several clusters and do not emerge from this analysis as separate research areas. 

Trends in the Content of Biomedical Research 

One straightforward application of our typology of biomedical research is an 
examination of trends in the content of biomedical research over the five-year period 
1971-1975. Most knowledgeable observers of biomedical research in the United 
States agree that it has been changing rapidly as the nation's health priorities have 
changed, as new methods of research have developed, and as new links have been 
found between basic life processes and the nation's identified health priorities. Our 
research clusters provide an objective confirmation of this opinion. Some types of 
biomedical research have grown at a dramatic rate during the five-year period being 
studied, while others have grown slowly or even declined in size during the same 
period. Table 43 identifies 16 of the 50 research clusters that have grown at an 
annual rate of 10 oercent or more during 1971-75. Cluster size here is defined as the 
number of projects in the cluster being supported by NIH each year (i.e., number 
of competing or continuation grants awarded). Radiation biology (cluster 47) heads 
the list, growing at an average rate of 28 percent per year. Population and social 
science studies (clusters 29 and 31) have experienced growth rates near 20 percent 
annually. Three immunology clusters (6, 7, and 9) have grown between 12 and 14 
percent annually. Two genetics clusters (39 and 45) also appear on the high-growth 
list. 

Table 44 shows seven clusters that have declined in size by 2 percent or more 
per year. These clusters seem to have little in common. The small number of clusters 



99 

Table 43 



CLUSTERS WHICH HAVE BEEN GROWING BY 

10 PERCENT OR MORE PER YEAR 

DURING 1971-75 



Cluster 



Growth Rate 
Per Year 



47 Radiation Biology 28 
29 Population Studies 22 
31 Social Sciences 19 
45 Molecular Genetics 17 

43 Biochemistry Using Biological Subsystems 16 

6 Nonclinical Immunopathology 14 

7 Immunology of Microorganisms 12 
9 Immunology (Lymphatic System of Animals) 12 

13 Animal Studies of Reproductive System 12 

27 Pharmacology Using Biological 

Subsystems 12 

39 Biochemical Genetics of Microorganisms 12 

42 Chemistry-Biochemistry 11 

44 Cell Biology and Biochemistry 11 
18 Surgery (Mostly Cardiovascular) 10 
26 Toxicology 10 

48 Biomathematics and Bioengineering 10 



Based on the regression: 

in (competing grants + continuation grants) = a + b*year 



Table 44 

CLUSTERS WHICH HAVE BEEN GROWING SMALLER 
BY 2 PERCENT OR MORE PER YEAR 
DURING 1971-75 



Cluster 



Growth Rate 
Per Year 



17 Nondrug Related Physiology of the 

Cardiovascular System -8% 

41 "Other" Chemistry -4 

21 Clinical Physiological Studies -4 
23 Drug Related Human Clinical Studies -3 
20 Hematology _2 
36 Chemistry of Biopolymers Using Animals -2 

22 Clinical Physiological Studies Using Animals -2 



Based on the regression : 

fin (competing grants + continuation grants) = a + b*y Pa r 



100 



on this list and the small rates of change reflect the fact that the total number of 
research projects funded by NIH has been growing during the 1971-75 period. When 
the 50 clusters are ranked by their rate of growth or decline over this period, the 
median cluster grew at a 6 percent annual rate. 

National Health Priorities and the Content of Biomedical 
Research 

Why have some types of biomedical research grown much more rapidly than 
average since 1971 while other types of research have declined? At least two compet- 
ing explanations can be proposed. One is that changes in the content of biomedical 
research reflect changes in the scientific opportunity for work in different fields. In 
some fields, it could be argued, research is stalled for lack of appropriate research 
methods, lack of theory, or the need for more basic information from other fields 
before advances can be expected. In other fields, it might be argued, the potential 
for scientific advance is much greater than average because of recent advances in 
theory or methodology. 

A second explanation for the greater growth of certain types of biomedical 
research is that the Congress, through differential funding of the Institutes of NIH, 
has established its own priorities for biomedical research based not on scientific 
criteria alone but also on the basis of the nation's health needs. The extent to which 
criteria other than scientific merit should be used in choosing among competing 
research applications has been, of course, a matter of considerable debate within the 
biomedical research community. Our present task is not to pass judgment on this 
question, but rather to ask some empirical questions that may clarify the debate. 
First, to what extent do priorities other than scientific merit actually play a part in 
research funding decisions? Second, is there any evidence that funding decisions 
based on national health priorities can influence the amount of high-quality work 
being done in a particular field? 

Each competing approved research application receives a priority score as part 
of its review by an NIH Initial Review Group (sometimes referred to as a study 
section). This group consists of approximately 12 to 15 scientists who are knowledge- 
able in the scientific area of the application. Each study section member assigns a 
score between 1 and 5 based on the scientific merit of the application. These scores 
are averaged to produce the priority score for the application. 

Suppose NIH were not NIH, but rather a different federal agency whose only 
concern was to fund biomedical research of the highest possible scientific merit 
without regard to its relevance to any specific disease problem. It might use the 
identical review procedure used by our current NIH and place all applications in a 
single list ordered by priority score — which is the only indicator of scientific merit 
available to the agency. It would then go down the list funding all applications until 
it had spent its budget authorization. 

How different would this system be from the way NIH operates now? We found 
that 84 percent of the decisions to fund or not to fund applications would have been 
the same if priority score alone had been the basis for funding. Going further, we 
calculated how many applications in each cluster would have been funded if a 
uniform priority score cutoff had been applied across all the Institutes. Six of the 
50 clusters have consistently received funds above what they would have received 
under a "priority score only" system, and seven of the 50 have consistently received 



101 

funds below what they would have received under a "priority score only" system. 
These clusters are shown in Tables 45 and 46. 

With the exception of social sciences (cluster 31) and pathology -oncology (cluster 
49), the minimum support level for research clusters that consistently received extra 
funding has been only modestly above the expected level. For most of the clusters 
that have been consistently underfunded (that is, clusters receiving fewer grants 
than would be predicted from scientific priority scores alone), the maximum support 
level has been only modestly below the expected level. An exception has been cluster 
39 (biochemical genetics with microorganisms), which, in its best year, received only 
89 percent of the grants it would have received if priority score had been the sole 
basis for grant decisions. 

The data in Tables 45 and 46 enable us to check, in a crude way, whether 
national health priorities (as reflected in consistently high or low funding levels for 
research clusters) have caused the amount of research in heavily funded clusters to 
increase or the amount of research in underfunded clusters to decrease. The cluster 
labeled "social sciences" has consistently been blessed with a high level of funding 
(relative to the priority scores on applications), and research in this area has grown 
dramatically during the 1971-75 time period. However, another heavily funded 
cluster (pathology-oncology) has grown only at a 3 percent annual rate, below the 
median; and most of the rapidly growing clusters in Table 43 have not consistently 
received heavy funding by NIH (again, relative to the priority scores on appli- 
cations). 

The same mixed pattern emerges when we look at clusters that have consis- 
tently received less funding than would be expected on the basis of scientific priori- 
ties. Two of the consistently underfunded clusters (chemistry of biopolymers using 
animals and chemistry of biopolymers using microorganisms) have also been declin- 
ing in size over the 1971-75 period. A plausible interpretation is that the shortage 



Table 45 



CLUSTERS WHICH HAVE CONSISTENTLY RECEIVED 
HIGH PRIORITY FUNDING 



Cluster 




Minimum 
Support Level 


49 


Pathology-Oncology 


1.27 


31 


Social Sciences 


1.17 b 


24 


Chemistry-Pharmacology 


1.07 


6 


Other Immunopathology 


1.03 b 


27 


Other Pharmacology 


1.03 


21 


Clinical Physiological Studies 


1.03 



Ratio of actual number of grants to expected number based on 
priority score. 

This has been a rapidly growing cluster. 



102 



Table 46 



CLUSTERS WHICH HAVE CONSISTENTLY RECEIVED 
LOW PRIORITY FUNDING 



Cluster 


Su 


Maximum 
pport Level 


39 


Biochemical Genetics with Microorganisms 


.89 b 


32 


Biochemistry and Genetics of 
Microorganisms 


.91 


40 


Chemistry of Biopolymers Using 
Microorganisms 


.92 C 


A8 


Biomathematics and Bioengineer ing 


.94 b 


9 


Immunology, Other 


.97 b 


4 


Immunochemistry of Blood 


.98 


36 


Chemistry of Biopolymers Using Animals 


.98 C 



Ratio of actual number of grants to expected number based on 
priority score. 

This has been a rapidly growing cluster. 

c 
This has been a declining cluster. 



of funds for research in biopolymers has depressed investigators' interest. However, 
three of the consistently underfunded clusters ("other" immunology, biochemical 
genetics with microorganisms, and biomathematics and bioengineering) have been 
growing rapidly. This suggests that the demand for research funding in these areas 
is pushing the supply, and that the supply of funds for these clusters, while growing, 
has never quite caught up with the increasing demand. 

This examination of the clusters that have grown most and have grown least 
does not provide a definitive answer to the question of whether the level of funding 
of research attracts or discourages researchers. We hypothesize that demand for 
research support in an area would depend on the scientific opportunities available 
in that area, the number of available scientists who possess the training and ability 
to work in that area, and the perceived likelihood of obtaining research support. We 
do not have any way of directly measuring any of these quantities, so we have 
attempted to develop surrogates for each one. 

It should be possible to develop a proxy for the kind of scientific opportunities 
available in an area from the distribution of the priority scores received by appli- 
cations in the cluster. The priority score is a measure of the scientific merit of an 
application and one would expect that in an area with many exciting opportunities 
for research projects, there would be more applications with better than average 
scores. We tried using the average priority score awarded to applications in the 
cluster, as well as the proportion of applications in the cluster with scores among 
the best 10 percent and 20 percent of all scores. However, these variables were never 
significant in our regression equations. It may be true that when new scientific 
opportunities arise, they attract new projects that are of low scientific quality in the 



103 

same proportion as those of the highest quality. In any case, we were unable to 
develop a proxy for scientific opportunity that was correlated with the growth in 
applications. We present our regression results without the variable from priority 
scores. Including this variable in the regression leads to identical conclusions. 

Since we are considering only a five-year time span, the number of qualified 
scientists in any one field probably has not changed too much over this period. 
However, the availability of that manpower may change with changes in support 
levels. If many grants are awarded in an area in one year, then at least those people 
are unlikely to reapply the next year. Thus we have also counted the number of 
continuation grants awarded in each cluster in each year. We expect changes in this 
variable to be negatively correlated with changes in number of competing appli- 
cations. 

For a proxy of the perceived likelihood of funding in a cluster, we use the 
difference between the actual number of grants awarded in the cluster and the 
number of applications with scores better than our hypothetical priority score pay 
line — i.e., the number of grants that would have been awarded if scientific merit as 
measured by priority score had been the only criterion used to fund applications. We 
would expect that if a larger number of grants is awarded in a research subfield than 
would be expected by priority score alone, then the number of applications for 
research support in this area would increase sometime later. 

To specify our model completely, we need to describe the lag structure. Table 47 
shows a regression of the number of applications in each year on the number of 
applications in the previous year, the change in number of continuations between 
the two years and the difference between the number of grants awarded in the 



Table 47 



REGRESSION OF APPLICATIONS PER CLUSTER BY YEAR 



1972 1973 1974 1975 

R 2 0.72 0.66 0.82 0.89 

Applications in year 

t-1 0.83 a 0.85 3 1.29 3 1.02 

(8.7) (8.9) (13.1) (15.6) 

Changes in continuations -0.58 3 -0.52 -0.60 -0.35 

(3.7) (2.8) (2.5) (2.0) 

Preference given to cluster 

in year t-1 -0.08 -0.18 -0.39 -0.35 

(0.2) (0.4) (0.8) (1.0) 

Constant term 19.8 13.5 2.32 4.33 



t-statistics given in parentheses 
significant at 0.01 level 
significant at 0.05 level 



104 



previous year, and the number of grants above our hypothetical pay line in the same 
year. The change in continuations has the expected negative sign, but the variable 
for program priorities is never significantly different from zero. It would appear that 
federal program priorities cannot affect the receipt of applications in just a single 
year. 

In order to look at the longer term effect of federal program priorities, we 
examined the number of applications in fiscal years 1974 and 1975 as a function of 
what happened in the cluster in 1971 and 1972. For the manpower working in the 
field, and therefore not likely to apply for a competing application in the two-year 
period (t, t + 1), we use continuation applications in year (t + 1) minus grants 
awarded on a competitive basis in year t. The other independent variables are the 
number of applications received in 1971 and 1972, the number of grants awarded 
in 1971 and 1972 minus the expected number of grants in that period. There is 
evidence here of a response by the scientific community to availability of funding. 
For every application awarded beyond the nominal cutoff line in a research subfield 
in 1971 or 1972, two applications were received in 1974 or 1975. Thus it would appear 
that although program priorities cannot affect requests for support in the very short 
term, they can affect requests for support over a longer period. 

Research Content by Institutional Setting 

The scientific content of the research that is performed in different parts of the 
university is another question that we can address. We consider four parts of the 
university: (1) basic science departments of medical schools, (2) clinical science 
departments of medical schools and hospitals with a major affiliation with a medical 
school, (3) graduate schools and schools of arts and science, (4) the other health 
professions schools of dentistry, public health, pharmacy, and nursing. The last 
category is aggregated because each component receives very few NIH grants. The 
entire category accounts for only 6.5 percent of the competing research project 
applications in our files. 

An analysis of the relationship between scientific subfield based on the clusters 
of applications and the components of the university shows that each component of 
the university plays a unique role in the spectrum of biomedical research. 

Most of the clusters of clinical studies are performed almost solely by members 
of the clinical science departments of medical schools (these clusters are 5, 9, 21, 22, 
23, 33). The exception to this rule is cluster 28, which is clinical development studies 
that are performed in all components of the university except for the basic science 
departments of the medical school. 

Two of the clusters identified by a medical specialty (surgery and radiology) are 
also performed almost entirely within the clinical department of the medical school. 

One can also identify several subfields that are performed almost solely within 
the medical school but in both basic science and clinical departments. These are: 
drug-related studies of the nervous system (cluster 1), drug-related pharmacology 
(27), drug-related biochemistry (25), drug-related physiology of the cardiovascular 
system (16), pathology, immunology, other (9), microbiology and immunology (8), 
immunopathology (6), biochemistry and physiology of the endocrine system (14), 
physiology of the cardiovascular system (17), and hematology (20). Thus it would 
seem that some of the research performed in basic science departments of medical 



105 



schools is sytematically different from research performed in graduate schools in 
being either drug-related, based on a body system, or related to a medical specialty. 

Much of the basic research within departments of medical schools is also per- 
formed within graduate schools but not often within clinical departments. These 
areas are several of the biochemistry, molecular genetics, and cell biology clusters 
(32, 36, 39, 40). One can also identify fields of biomedical research that are performed 
in the university departments and hardly ever in medical schools. These areas are 
the chemistry clusters (24, 41, 42). Only the chemistry and pharmacology cluster is 
drug related. 

Clinical departments of medical schools and university departments perform 
research not performed in basic science departments of medical schools only in the 
clusters within the behavioral sciences (28, 29, and 30). 

Figure 4 summarizes the research areas in which each of the components of the 
university specializes. The clusters that have not been mentioned above receive 
applications from each part of the university approximately in proportion to total 
applications. 



CONCLUSION 

It is possible to use the scientific classification code system developed by the 
Division of Research Grants to produce a detailed typology of biomedical research 
projects. We have produced a preliminary set of clusters of applications for tradition- 
al NIH research project grants from institutions of higher education. 

The rate of funding of approved NIH applications varies depending on the area 
of research of the application but does not vary by component of the university. On 
the average, applications for research support in a scientific field will increase for 
several years following an influx of NIH support for that area, but will not respond 
to year-to-year changes. The typology of research projects has also been used to 
describe differences in the research performed in different components of the univer- 
sity. 



-106- 



University Basic 
Science Departments 



Medical School 

Basic Science 

Departments 



Chemistry 



Biochemistry 
Molecular 
Genetics 
Cell Biology 



Medical School 
Clinical Science 
Departments 



University 
Behavioral Sciences 
Departments 




Most Drug Related 
Studies 

Organ System 
Studies - Not 
Using Human 
Subjects 

Subset of 

Microbiology 
and Immunology 



Most Studies 
Using Human 
Subjects 

Surgery and 

Radiology 



Psychology 
Behavioral 

Science 
Population 

Studies 



Fig. D-l Research specialization by institutional setting 



VI. GENERAL CONCLUSIONS 



Our analysis has focused on specific questions regarding the effects of federal 
biomedical research programs on academic medical centers. However, the purpose 
of the overall study is to provide a broader understanding of the degree and the 
determinants of the interdependency between the federal agencies that sponsor 
biomedical research and the academic institutions that perform it. Because the 
efficiency of federal biomedical research programs greatly depends on the efficiency 
of these nongovernmental institutions, this interdependency makes the status of 
these academic institutions a matter of federal concern. 

Our analysis addresses the question of the status of the academic medical com- 
munity on a number of levels. It examines how centers appear to have adjusted their 
educational programs, their organizational structures, their scientific activity, and 
their budgets as a result of their involvement in federal biomedical research. 

The measurable effects of federal research on the educational programs of cen- 
ters appear to be limited largely to those components most involved in research. We 
observe that federal research training programs specific to a scientific field and the 
overall research intensity of the institution are the factors that seem best to explain 
the size of Ph.D. programs in the basic science departments. Moreover, there is 
limited evidence to suggest that these departments significantly reduce enrollments 
when training grant funds are cut back, but the effects of these cutbacks may be 
mitigated in departments with substantial research funding. 

By contrast to the situation in basic science departments, we do not observe 
strong effects of federal biomedical research on the major educational programs of 
clinical departments, and this is consistent with their major involvement in the 
delivery of patient care in conjunction with education. Our analysis of the size of 
graduate medical education programs in internal medicine suggests that the num- 
bers of interns and residents are determined largely by the patient loads in teaching 
hospitals and by the size of the clinical faculty. The intensity of a clinical depart- 
ment's research activities does not appear to have a significant effect on the size of 
its house staff programs. We also find that the research intensity of a medical school 
has very little or no effect on the specialty choices of its graduates. Not surprisingly, 
M.D. graduates with the best academic records are the most likely to enter academic 
or research careers, and this likelihood is increased if they attended a research- 
intensive medical school. However, only a small proportion of the graduates of even 
the most research-intensive medical schools enter academic and research careers. 

Departments are the most important organizational units of academic medical 
centers, and our analysis indicates that their success in the competition for federal 
research funds has an important effect on their size. This effect is most apparent in 
the case of basic science departments. The less strong apparent effects of research 
on clinical departments may be due to the limitations of data to account for patient 
care activities. 

The total budgets of all departments appear to be quite sensitive to the rise and 
fall of their federal research funding. Moreover, substantial proportions of the 
salaries of research oriented faculty come from federal grants, indicating they may 



107 



108 



be vulnerable to federal funding cutbacks. Although the research funds are impor- 
tant to many departments, the research intensity of a department is negatively 
related to salary levels of its members, when local economic conditions are con- 
trolled for. 

If an academic medical center is substantially involved in biomedical research, 
it must depend heavily on federal funds for budget stability. Our analysis of all 
funding sources indicates that federal research funds do not appear to attract sub- 
stantial funds from other sources. Similarly, centers have little flexibility to compen- 
sate for federal research funding cutbacks with funds from other sources. This is 
consistent with our observation that deans generally do not use institution funds to 
"bail out" departments that lose research funding. 

In examining the scientific characteristics of federally sponsored biomedical 
research in academic institutions, we find strong evidence of specialization along 
both federal program and scientific lines. The different Institutes of NIH and 
ADAMHA depend on different organizational components of medical centers and 
universities. This means that a particular set of departments, or a particular class 
of institutions, may be quite vulnerable to funding cutbacks by a single Institute, 
while many others may be hardly affected at all. Similarly, the success of the 
research programs of a particular Institute may be quite sensitive to the situation 
of a particular segment of higher education. 

An individual investigator's research area is largely determined by his training 
and prior experience. However, there is clear evidence that the true scientific char- 
acteristics of research proposals are influenced by federal program emphasis as 
measured by the likelihood of funding among scientific groups. This suggests that 
the federal government has the capacity to affect not only the level but the nature 
of scientific activity within the biomedical research community. 

In the most general terms, the results of our analyses appear to describe an 
academic medical community that is responsive to the influence of federal biomedi- 
cal research programs. They tend to confirm the interdependence between the fed- 
eral agencies that sponsor research and the academic institutions that perform it. 
Important characteristics of academic medical centers — Ph.D. programs, faculty 
size, budgets, scientific activity — are directly related to the federal funding received 
by individual departments; and the overall financial stability of research-intensive 
centers is substantially affected by the stability of federal research funding. 

It is significant that centers must make long term commitments — hire new 
faculty, alter research directions, admit students — in order to respond to federal 
research policy that appears increasingly subject to short term shifts. In the past, 
centers have undertaken long term commitments to respond to federal programs, 
only to have the particular programs deemphasized by the government. However, 
in recent years, the accommodation to these unexpected changes in federal program 
emphasis has been eased because the centers have been able to devote these re- 
sources to other activities that were part of the growing federal program involve- 
ment in academic medicine — expanded M.D. enrollment, Medicaid, cancer research, 
family practice training, allied health professional education, and so on. We see little 
evidence in recent policy debates to suggest that the federal program involvement 
with academic medical centers will continue to grow at rates approaching those of 
the past. Hence, we would expect that any future short term shifts in federal 
biomedical research policy will have much more adverse effects on the academic 
medical community than such shifts have had in the past. 



PART V 



A Report on the Indirect Costs of Academic Research 



American Council on Education 
Association of American Medical Colleges 
The Rand Corporation 



TABLE OF CONTENTS 

Page 

I. Introduction 1 

II. Procedure 2 

III. The Determination of Indirect-Cost Rates . - 4 

IV. Policy Limitations upon the Recovery of Costs of Research 5 

V. Rising Trends in Indirect Costs 7 

VI. Indirect-Cost Problems Arising within Academic Research 

Institutions 9 

VII. Proposals for Revisions of Federal Indirect-Cost Regulations.... 11 

VIII. General Conclusions 14 

References 17 

Appendix A 18 



-l- 



A REPORT ON THE INDIRECT COSTS OF ACADEMIC RESEARCH 



I. Introduction 

A study of the policy implications of the indirect costs of academic 
research was one of the tasks specified by the President's Biomedical Research 
Panel in the contract with the American Council on Education for "Studies of 
the Impacts of Federal Health-Related Research Expenditures upon Institutions 
of Higher Education." The Panel defined its interests in the problem of indirect 
costs in terms of the following questions: 

1. How are indirect costs determined and indirect cost rates negotiated? 

2. How is the audit process conducted for renegotiation of indirect cost 
rates? 

3. What differences are there between federal and non-federal funding 
agencies in policies covering indirect costs and the development of 
indirect-cost rates? 

4. What patterns exist in the control and allocation of federal indirect- 
cost reimbursements? 

5. What policies and restrictions are followed as to the uses that can be 
made of such funds? 

6. What is the institutional attitude toward present federal percentage 
levels, federal restrictions, and federal administrative procedures 
relative to such funds? 

It was agreed in the proposal for the general investigation that the study 
of indirect costs would be the joint responsibility of the American Council on 
Education (ACE, the prime contractor) and its two subcontractors, the Association 
of American Medical Colleges (AAMC) and the Rand Corporation (RAND). The present 
report has been prepared, therefore, under the joint aegis of the project direc- 
tors for the three organizations: Lyle H. Lanier (ACE), Thomas E. Morgan (AAMC), 
and Albert P. Williams (RAND). 



-2- 



II. Procedure 

It was decided that systematic answers to most of the Panel's questions 
could best be provided through the preparation of a comprehensive monograph by 
an individual with extensive experience in university research administration 
who had also been involved in the participation of higher-education representa- 
tives in the development of federal indirect-cost regulations. Raymond J. 
Woodrow of Princeton University was selected for this task as being the best- 
qualified person available. His monograph entitled Indirect Costs in Univer- 
sities (11) is being submitted as a supplement to the present report; and the 
Table of Contents has been reproduced herein as Appendix A. 

Other contributors of material for this study included Frederick B. Putney 
of Columbia University who prepared a paper elaborating on the numerous miscon- 
ceptions of the nature of indirect costs and providing a detailed example of 
indirect-cost calculations for a hypothetical university. 

On a different aspect of the problem, George W. Baughman of Ohio State 
University assembled data designed to explain why indirect costs have been 
increasing recently at a more rapid rate than the direct costs of research. 
His source was the Ohio Higher Education Price Index of the Ohio Board of Regents, 
which is based on price-change data for the twelve state-assisted universities 
in Ohio. Comparative percentage changes for selected components of this 
price-index series have been included in this report. 



Raymond J. Woodrow served for many years as Director of Research Administration 
and Executive Secretary of the University Research Board at Princeton University. 
He was a consultant and member of various national committees concerned with 
federal indirect-cost regulations, and has published several articles on the sub- 
ject (cited among the references in his present monograph). Since his retirement 
from the position named above, he has directed an NSF-supported project on research 
management in universities under the aegis of Princeton's University Research Board. 



-3- 



Inasmuch as proposals for substantial changes in federal regulations gov- 
erning reimbursements for the indirect costs of academic research have recently 
appeared, a final section has been included in this report on the status of 
these developments. Particular attention has been given to the communications 
between representatives of universities and their associations and representa- 
tives of the Executive Branch and the Congress for the purpose of achieving 
agreement on procedures for resolving the issues raised. 

The following sections consist mainly of digests of material from Woodrow's 
monograph and other sources that relate to the Panel's questions and to other 
aspects of the policy implications of the indirect costs of academic research. 
The discussion has been organized under six headings: (a) the determination 
of indirect cost rates; (b) policy limitations upon the recovery of costs of 
research; (c) rising trends in indirect costs; (d) indirect-cost problems arising 
within academic research institutions; (e) recent proposals for revisions of 
federal indirect-cost regulations; (f) general conclusions. 



-4- 



III. The Determination of Indirect-Cost Rates 

Federal policies and procedures governing reimbursement for the indirect 
costs of research conducted by universities have evolved through numerous stages 
since 1947 , when the first set of regulations provided for the determination of 
an average indirect-cost rate covering all instructional and research activities 
of an institution — based largely on its regular financial report. They were 
superseded in 1958 by Bureau of the Budget Circular A-21, which was applicable 
only to research costs and which established a systematic set of costing prin- 
ciples and general guidelines (but not a detailed set of uniform accounting 
procedures) . The document was developed by an interagency committee which 
worked closely with a group of university representatives organized by the 
American Council on Education. 

Circular A-21 has been revised five times. In 1973, the responsibility 
for this function was transferred to the General Services Administration — which 
reissued the regulations without substantial change as Federal Management Cir- 
cular No. 73-8 (FMC 73-8). Recently, the Office of Management and Budget (OMB) 
has reassumed this administrative responsibility. 

Several points regarding this history of the federal regulations governing 

indirect costs and indirect-cost rates are pertinent to the major issued raised 

in discussions of this subject; 

1. The research activities conducted by universities generally require 
numerous services and other support which fall under the accounting 
category of indirect costs and which can be most equitably prorated 
to individual projects on the basis of some such principles as those 
embodied in FMC 73-8. 



-5- 



2. Indirect costs are real costs that must be met by an institution from 
its operating budget. If individual projects or other activities 
requiring indirect-cost services do not pay their pro rata share of 
the cost, it must be borne by the remainder of the institution's 
budget. 

3. Differences of opinion exist between federal agencies and universities 
as to what indirect costs are allowable as charges against research 
grants, but the responsible federal agency makes decisions in these 
cases and they are reflected in the indirect-cost rates finally ap- 
proved after an appropriate auditing process. 

A. To a considerable degree, current efforts to change the indirect- 
cost regulations arise from unsatisfactory accounting practices and 
poorly documented requests for reimbursement on the part of some 
institutions that presumably are attempting to comply with the regu- 
lations and the terms of their audited indirect-cost rates. Many of 
these criticsms are undoubtedly justified, and the higher-education 
associations mainly concerned are taking steps to urge their member 
institutions to institute more rigorous accounting and documentation 
procedures in claiming indirect-cost reimbursements. 

5. The efforts to revise the policies and regulations, however, go far 
beyond steps to assure better compliance by a minority of institu- 
tions. The proposed changes would limit further the range of in- 
direct costs eligible for reimbursement under research grants and 
contracts; and they would institute enormously costly changes in 
detailed accounting and reporting procedures (which the present FMC 
73-8 specifically views as undesirable). 

IV. Policy Limitations upon the Recovery of Costs of Research 

The Congress and the Executive Branch have from time to time imposed 
ceilings or other forms of limitations upon full recovery of the costs of 
research by academic institutions. Such limitations have usually applied to 
grants rather than contracts (the limitations to such recovery under contracts 
relating mainly to what charges are "allowable" as indirect costs of research). 
In the case of research grants, initially there was a period in which indirect- 
cost reimbursement was limited to a fixed percentage of direct project costs. 
This percentage for NIH, for example, was first established at 8 per cent of 



-6- 



direct project costs — a limit that was first raised to 15 and later to 20 per 

cent for HEW projects. Curiously, independent agencies, by contrast, were 

allowed a reimbursement limit of 25 per cent for indirect costs for one year; 

but this was reduced to 20 per cent for the following year. 

The Congress in 1965 abolished all indirect-cost ceilings but replaced 

them with a policy of "cost sharing." This statutory restriction provided that 

no recipient of a federal grant for research should be paid "as much as the 

entire cost of the project." Individual agencies were left reasonably free to 

determine how the cost-sharing policy should be implemented. 

(Since 1969, cost-sharing has also been required on the contracts of 
certain federal agencies, other than DHEW, which result from proposals 
not specifically solicited by the agencies. The extent of cost-sharing 
in such cases is supposed to reflect the respective degrees of institu- 
tion and of agency interest in the research.) 

One reason given for requiring universities to share in the costs of 
grant-supported research is that such activity is an integral part of the in- 
stitution's regular educational program and hence should not impose an extrane- 
ous added burden. In the abstract, this argument is a plausible one — and 
it is similar to the one usually made by foundations — but conformity to its 
"logic" would most assuredly curtail the level of university participation in 
the nation's research effort or else would seriously distort the total educa- 
tional programs of universities by siphoning off resources required for the 
effective performance of important "non-sponsored" activities. The deteriora- 
ting general financial situation of research universities, caused by the com- 
bined effects of inflation and recession, makes the burden of sharing in the 
costs of federal programs increasingly difficult for institutions to bear. 



-7- 



Since precise information as to the full degree of cost sharing and as 
to the nature of its impacts upon different types of institutions is unavail- 
able, it is believed that a systematic study of this problem should be spon- 
sored by an appropriate government agency or agencies. Consideration should 
be given in such a study to a comparison of the federal indirect-cost policies 
applicable to industry with those prevailing for universities. 
V. Rising Trends in Indirect Costs 

The evidence is reasonably clear from several sources that the indirect 
costs of research as a proportion of total costs (or as a percentage of direct 
costs) have been rising in recent years. The evidence is not so clear, how- 
ever, as to the magnitudes of these increases for various types of research 
activities or among different institutions. Furthermore, the reasons for 
such increases are not fully understood or adequately appreciated by many who 
are concerned about them. The present summary will present evidence on the 
trends in such increases for NIH grants to universities in the ACE sample, and 
will then cite price- increase data to show why indirect costs have increased at 
a faster rate than direct costs. 

1. Changing Proportions of Indirect-Cost Funds in NIH Research Grants . 
From the NIH IMPAC files, records of the grant funds awarded to 145 
universities were secured for a separate ACE study (6) , together 
with the percentages of these funds provided for direct and indirect 
costs. Two types of trend indices have been computed for indirect 
costs: (a) the proportion of the total awards allocated for indirect 
costs; (b) indirect costs as percentages of direct costs. The 
following is a tabulation of these two sets of percentages covering 
the fiscal years 1969 through 1975: 

1969 1970 1971 1972 1973 1974 1975 

Indirect Costs „„ „„, 

Total Costs 21-7% 23,5% 24 ' 4% 25 ' 6% 26 ' 8% 27 - 5% 26 - 5% 

Indirect Costs „-, -, „„ ., 
Direct Costs 27 ' 7 30 - 7 32 ' 3 34 ' 4 36 " 6 37 " 9 36 '° 



-8- 



Comparative Increases in Prices for Direct and Indirect Costs of 
University Operations . There are several reasons why indirect costs 
have increased more rapidly than direct costs of research in recent 
years. Probably the most significant has been the differential 
effects of inflation on the two types of expenditures. The direct- 
cost component of research expenditures has a higher proportion of 
salaries and wages than the indirect-cost component; and the former 
has been increasing at lower rates than the latter. The price in- 
creases for books and periodicals, utilities, and other non-personnel 
items have escalated far more rapidly in recent years than personnel 
compensation (which is the denominator in most formulae for deter- 
mining indirect-cost rates) — as the following price-change percentages 
show: 

AAUP*^ Ohio Higher Education Price Index 



Fiscal 


Faculty Com- 
pensation 

7.2% 


All 
Prices 


Non- 


-Personnel 


Prices 


Year 


All 


Library 


Phys. Plant 


1968-69 


5.7% 


3.0% 


4.8% 


2.8% 


1969-70 


7.1 


6.8 


5.9 


10.6 


8.6 


1970-71 


6.2 


6.0 


8.5 


11.5 


9.7 


1971-72 


4.3 


4.4 


4.8 


9.6 


5.0 


1972-73 


5.0 


5.9 


9.0 


12.1 


10.2 


1973-74 


5.9 


9.0 


18.6 


13.3 


29.7 


1974-75 


6.4 


10.6 


14.8 


18.2 


21.0 



Another major factor contributing to rising indirect costs has been the 
costs of compliance with federal laws and regulations — e.g., those relating to 
such federally mandated social programs as the following: equal employment 
opportunity, equal pay, affirmative action, occupational safety and health, 
minimum wage and fair-labor standards, Social Security increases, health-main- 
tenance organizations, and environmental protection (10). 



'See References, number 1. 



Provided by George W. Baughman of Ohio State University from the price indices 
developed by the Ohio Board of Regents from the records of the twelve state- 
assisted universities in Ohio. 



-9- 



More specifically affecting federal grants and contracts are regulations 
relating to the utilization of labor surplus, small business concerns, minority 
business enterprises, the use of human subjects in research, animal-care require- 
ments, employment openings for veterans, and employment of the handicapped. 
Compliance with these regulations, as well as with the increasingly detailed 
information requirements under grants and contracts, all have increased the 
indirect costs of the administration of sponsored research. 
VI. Indirect-Cost Problems Arising within Academic Research Institutions 

Two general types of intra-institutional problems related to indirect costs 
may be identified: (a) those relating to the determination of the indirect- 
cost rate structure for the institution, particularly as regards complex uni- 
versities with large and varied types of organizational units; (b) and those 
affecting the internal allocation and use of indirect-cost reimbursements 
received by the institution. 

The problem of determination of the appropriate rate structure for indirect 
costs at an institution is the responsibility of the federal agency in charge 
of review, audit, and establishment of rates under federal regulations. Separate 
indirect-cost rates might be established within a complex institution for 
large and relatively autonomous research units whose indirect costs varied sub- 
stantially from those for the remainder of the institution. Such a determination 
presumably would be made in negotiation with the institution in terms of an 
equitable balance as between the institutional and the federal interest. 



-10- 



Conflicting interests and points of view often arise within research univer- 
sities regarding both the validity of indirect-cost charges and the equity of 
internal budgeting of indirect-cost revenues. Faculty members and department 
heads, for example, sometimes feel that their individual projects do not require 
all of the indirect-cost services that are specified in the negotiated rate 
structure. They fail to understand that the latter is based upon an averaging 
of the costs of such services for all projects at the institution (or within 
the part of it covered by a particular rate structure) . For the general execu- 
tive officers of the institution, on the other hand, the totality of indirect 
costs represents obligations that must be met; and such costs increase gener- 
ally in proportion to the expansion of research and other activities. Admin- 
istrators are concerned, moreover, with the problem of meeting the indirect 
costs that are not recovered under federal regulations and under the law re- 
quiring the sharing of the costs of grant-supported research. 

Steps need to be taken to assure better mutual understanding of the respec- 
tive attitudes and concerns of both faculty and administrative groups. It is 
especially important for faculty members and department heads to understand 
that if the costs of sponsored projects are not fully reimbursed, they must be 
met by the reallocation of institutional funds from other programs and purposes. 

The federal interest in the internal budgeting of indirect-cost revenues 
would appear to relate solely to the question of whether or not its sponsored 
projects have been provided with adequate indirect-cost services. It is not 
the federal responsibility or prerogative to enter into an evaluation of the 
educational budget of an institution, which includes indirect-cost revenues 



-li- 



as a component that must be integrated into the totality of the general revenues 
available for the educational and support operations of the institution as a 
whole. Federal responsibilities would seem to end with the determination that 
an institution's indirect-cost rates are justified, that its requests for re- 
imbursement are valid, and that the indirect-cost services required for feder- 
ally sponsored projects are satisfactorily provided. 
VII. Proposals for Revisions of Federal Indirect-Cost Regulations 

Two federal reports issued during 1975 have stimulated renewed discussion 
of reimbursement for the indirect costs of academic research. The first was 
submitted to the Committee on Appropriati